Re: Resizing LVM issue
On 6/14/2014 4:33 PM, Miroslav Skoric wrote: ...may I suggest to LVM programmers to think about some software routines that would enable users to recompose (resize, shrink, whatever ...) their LVM from within a mounted system, in a way that after the next reboot, the LVM and FS automatically recomposes itself - so to avoid common mistakes. This is not possible. A filesystem must be shrunk before the underlying storage device. If you shrink the LV first then portions of the filesystem will now map to non-existent sectors. If files exist in those sectors they will be lost. Same goes for filesystem metadata. It is possible to add sectors to a device under a mounted filesystem because the filesystem has no knowledge of them, and is not mapping them. The same is not true of removing sectors under a mounted filesystem, for the reason above. Cheers, Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/539cd28a.20...@hardwarefreak.com
Re: Large File Systems - Enough inodes?
On 5/20/2014 12:00 PM, Richard Hector wrote: On 21/05/14 04:24, Sven Hartge wrote: ... I like to create filesystems relatively small, on LVM, so that any of them can be grown later, when I find out where the space is needed. But extending an ext(2|3|4) filesystem doesn't create new inodes, so the ratio of inodes to space drops, and eventually this is a problem. And if you really want to be on the safe side: use XFS. And that's my solution. The reason for this is two fold. First, xfs gives you plenty of inodes to begin with, and xfs_growfs adds more inodes as well as additional free space when you grow an LV. Example using mkfs.xfs defaults: FilesystemTypeSize Used Avail Use% Mounted on /dev/sda6 xfs 94G 6.4G 87G 7% /home FilesystemTypeInodes IUsed IFree IUse% Mounted on /dev/sda6 xfs 94M7.1K 94M1% /home 1 million inodes per gigabyte. Cheers, Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/537bcf91.7050...@hardwarefreak.com
Re: Random hard freezes Wheezy
On 5/17/2014 2:43 AM, Mimiko wrote: On one server I intermittently encounter hard freezes. Server does not react, ping, or ctrl+alt+del. Just caps lock and num lock flashes. Only a power off from button helps to start server, after which it runs until again this happens. Flashing KB LEDs is typically a sign of hardware failure, either permanent or temporary. Usually this is a symptom, not a cause, but check the PS2 connection to the motherboard just in case. Make sure it is secure, not loose or wobbly, and make sure none of the 6 pins are bent or broken. ... PRIMERGY Econel200/D2020, BIOS 08.10.Rev.1100.2020 06/01/2006 This Fujitsu server is 7-8 years old... What could be the problem of this? Eth1 link down is likely a symptom, not a cause. However, it could be a cause if the switch port on the other end of the cable is going bad. In that case the switch port could be applying spurious voltage to the wire, which could cause this server to lock up. This is rare but I have seen it in the past. A short in the cable may cause this as well, but again this is rare. Cables are cheap, so replace it just in case. However, the fact that it runs for many hours between lockups suggests the cause is a thermal problem. Check all fans in the system to make sure they're spinning at full RPM and are free of dust buildup. This includes the CPU fan(s), chassis fans, and fan inside the PSU. Given the age of this machine, I'd simply replace every fan in it for good measure. If all the fans are clean and spinning at full RPM, the next likely cause is a bad PSU. Check the output voltage of all PSU rails with a voltmeter to ensure they are within specification. If you do not have the test equipment or if this is beyond your ability, take the machine to a qualified repair shop and have them check it. Or, as PSUs are relatively inexpensive, simply replace the PSU due to age as well as the lockup issue. The host name srv75 suggests a server farm. If you are tasked with maintaining a farm, I'd assume you have the requisite hardware background to perform this testing, troubleshooting, and repair. Cheers, Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/5377ef16.2060...@hardwarefreak.com
Re: Kernel fails to detect internal hard drives after routine apt-get upgrade
On 5/11/2014 11:17 PM, O wrote: Hi Stan, The output from dmesg is long. From within initramfs, I cannot mount usb drives, and I cannot seem to scp or ssh. So far, I have not been able to find a way to get the output from dmesg (from within initramfs) onto another file system so that I can post it here: any ideas? Boot the working kernel and dump the dmesg output for us. I'm not looking for errors with the point release kernel relating to the boot problem. I'm looking for exactly what hardware we're dealing with and how it's connected to the system. It is atypical for a kernel point release to bork an ATA or SCSI controller driver to the point the system won't load. Thus I'm guessing there is something 'unique' about your setup. dmesg should tell me that, as well as other needed information. Cheers, Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/53706ed3.6000...@hardwarefreak.com
Re: Kernel fails to detect internal hard drives after routine apt-get upgrade
Dunno if you saw this or not. Selim identifies the source of the problem below and possible fixes. Read on. On 5/12/2014 5:52 PM, Selim T. Erdogan wrote: O, 12.05.2014: Hi Stan et al., Booting from the working kernel, I have dumped dmesg here: http://pastebin.com/MBTDfgc4 I tried to save dmesg booting under the 3.2.0-4_amd64 kernel from within initramfs, to no avail (I cannot mount usb drives to save the information, and it does not see the network). However, when I added debug to the kernel line in the boot command, I was at least able to see the system messages while the errors were happening. Here is the relevant block of text, and sas is involved: ata7: sas eh calling libata port error handler sas: sas_ata_hard_reset: Unable to reset I T nexus? sas: sas_ata_hard_reset: Found ATA device sas: sas_ata_hard_reset: Unable to soft reset sas: sas_ata_hard_reset: Found ATA device ata7: reset failed (errno=-11) retrying in 10 secs Searching the web for Unable to reset I T nexus led me to this thread: http://www.gossamer-threads.com/lists/linux/kernel/1912604 ... which was posted a short time ago and appears to be the identical problem. However, I am struggling to understand what I should do in response. It seems to be saying that my hardware and its drivers are too new for Wheezy, even though this machine is 2 yrs old?? Does this mean I have to upgrade to Jessie? I happened to notice the following bug report while updating last week: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=746642 Basically, it seems like people solved this by booting from a rescue disk and downgrading to an older kernel. After you fix your system, I recommend installing the apt-listbugs package. That's what showed me the bug report while updating. The problem is a patch/commit added in 3.2.57-3 meant to fix one problem but caused another more serious problem--unable to boot or register the drives. Since you can boot an older kernel, there is no need to use a rescue CD. Simply boot the older kernel and manually install the latest 3.2.x available prior to 3.2.57-3, using apt or aptitude. $ aptitude search linux-image will show your the kernel versions available in your configured repositories. Cheers, Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/53716d8e.4050...@hardwarefreak.com
Re: Kernel fails to detect internal hard drives after routine apt-get upgrade
On 5/11/2014 8:44 PM, O wrote: Dear debian-users, A first-time post for me: I've never had a problem that was so serious. I've had Wheezy installed for many months now with default kernel 3.2.0-4-amd64. Following a routine apt-get upgrade, the kernel cannot detect any of the internal drives. After many *ata#: reset failed, giving up* and *udevd: timeout* errors, I am dropped to an initramfs Busybox prompt that cannot see any drives (fdisk -l sees nothing). Interestingly, I *can* boot into the old Squeeze kernel (2.7.x). From booting under the old kernel, I have tried: # apt-get remove linux-image-3.2.0-4-amd64 # apt-get install linux-image-amd64 # cd /boot; update-initramfs -k 3.2.0-4-amd64 -u # apt-get update # apt-get upgrade # apt-get dist-upgrade All to no avail. Interestingly, even the Debian 7.5 netinst CD will not see the drives, and asks me to select a driver (it makes no suggestions). When I drop to a prompt in recovery mode (another BusyBox prompt), fdisk -l does not see any drives. There is nothing exotic about my drives: they are four Western Digital 2 TB internal hard drives. This one has me stumped. I hope someone can help. Please let me know if there is additional information I should post, or if there is somewhere else that I should be posting instead. I would appreciate any leads. Full dmesg output would be far more helpful than your narrative. Please paste it inline so we can cut the irrelevant parts from our replies and highlight the problem parts. Cheers, Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/53703107.7060...@hardwarefreak.com
Re: Disk heads won't park [pat II]
On 4/29/2014 1:20 PM, Nuno Magalhães wrote: Reviving this thread since i tried turning the machine on again (and amybe another thread will bump this one). And, again (well i wasn't expecting it to go away), as soon as the machine starts - right after POST, even before GRUB - the drive starts making reading noise (like when an antivirus is scanning or the system is thrashing). The only way it stops is with hdparm -y (no wonders there). The drive isn't failing, but has failed. Replace it. Mechanical drive platters have hard coded track markers. These are created by a low level format at the factory on today's drives. Those with experience going back to MFM/RLL days may recall performing low level formats due to stepper motor issues. This was done by entering g=c800:5 in a debugger, which loaded and executed the controller's firmware format utility. Drive firmware reads the low level track markers in order to properly position the read/write head on user data tracks. The noise you are hearing is the head seeking across the platter trying to locate the track markers and it is unsuccessful. The likely cause is a worn out actuator return spring. This is a rather common failure mode. The cost of the return spring is about USD 0.0025, about a quarter of a cent. If the spring tension deviates too far from spec the head will no longer align to a track marker. While Q.C. is typically good on spring manufacturing, they are made from large spools of drawn steel wire. A slight imperfection in a few feet of a 10,000 foot spool will yield a few dozen springs that may not last long, in your case about 2.1 years. Worn out spindle bearings can also cause head locating problems, but in this case you'll usually notice a vibration in the PC case and likely a hum accompanying it. Q.C. on the bearing assemblies is much higher than return springs. The spindle assy is the most expensive part in a disk drive due to the manufacturing tolerance and spin balancing required. You usually don't see bearing wear issues until 5+ years of power on duty. And of course bearing wear is inversely proportional to spindle speed. I.e. 5K drive bearings should normally last longer than 15K drive bearings. No software tool can identify the cause of your problem. However, SMART has been telling you for some time that the drive was experiencing seek errors. Seek errors indicate a head positioning problem. A head positioning problem normally indicates a worn return spring, bearings, or possibly a problem with the voice coil or its drive circuit, though the latter is rare. Cheers, Stan This is a Seagate Barracuda ST31000528AS drive with a CC49 firmware upgrade. Here's a few other commands i tried: ... 7 Seek_Error_Rate POSR-- 075 060 030-35143152 ... I do assume it is failing, but i'd like to know why and which values are really tell-tale (for instance the WHEN_FAILED column above is empty, so i can't realyl draw any conclusions). This is a recently installed, headless system with almost nothing installed. Thanks, Nuno -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/535ffbc4.1060...@hardwarefreak.com
Re: Disk heads won't park [pat II]
On 4/29/2014 6:13 PM, Nuno Magalhães wrote: On Tue, Apr 29, 2014 at 8:21 PM, Stan Hoeppner s...@hardwarefreak.com wrote: The drive isn't failing, but has failed. Replace it. I already have another one on the way. I was going to buy Samsung but then learnt their drive division was bought by Seagate (which also bought Maxtor, the brand of the oldest drive on my desktop). I settled for a Toshiba DT01ACA100 which would've been here already if the store hadn't handed me an DT01ABA100 instead (rpm difference). Maybe i just got unlucky with this Seagate, we'll see. I assume the return spring repair would be both infeasable and way beyond USD 0.0025 or the cost of the new drive. :) Alas, such is the market. The weird thing is, to an extent, the drive kinda works/ed (i guess only one platter is damaged but i won't play expert). It'll make a lovely paper-weight, though. It's not feasible to effect repairs to the moving parts of a modern hard disk drive. To do so would require breaking the air seal, and doing that will introduce dust particles into the platter cavity. Screw the cover back on and fire it up, and in no time flat the dust particles will scour the platter surfaces, as they get bounced around at 5900 to 15,000 RPM, and if one hits a read/write head it will damage it. Some of the professional data recovery services have clean room facilities and trained personnel (or used to anyway) who can effect such repairs in herculean efforts to recover data, but the drives are never returned to service. Last I heard such services start around $10,000 USD with no guarantee of data recovery from the failed drive. No software tool can identify the cause of your problem. However, SMART has been telling you for some time that the drive was experiencing seek errors. That's a subtle hint for me to setup smartd for the other drives (it was planned... in the to-do stack). Don't bother. Cron 'smartctl -A /dev/[device]' the first of every month and have it mail the output to you. Look at the raw values for seek and read error rates, reallocated sectors, etc. Once these pass zero and keep climbing it's time to get a replacement on the way. At that point the drive may have a year left, maybe a month, or maybe it will die tomorrow. No way to know. Don't wait until a drive is completely dead before replacing it. That mentality is for kitchen appliances, not something that won't spit your data back out after it goes ker-thunk. ;) Thank you for your very thorough explanation. You're welcome. Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/536059f3.9070...@hardwarefreak.com
Re: SMART data, should I change my HDD?
On 4/28/2014 5:43 PM, KS wrote: I was checking one of my systems and the SMART data for /dev/sda came out as below. Should I change it to avoid loosing data? If not, which information in SMART data indicates that it is time to do it? This drive is fine. 42C is quite warm for a drive but within acceptable range for the Caviar Black series. Max sustained operating temp is IIRC 65C, with a max short term peak temp of 80C. Below are the critical parameters that inform you of surface defects, spindle bearing wear, problems with the voice coil actuator hardware or electronics, and the external interface, i.e. SATA. Your raw values for all of these is zero, which means the drive is operating perfectly. ... 1 Raw_Read_Error_Rate 0x002f 200 200 051Pre-fail Always - 0 ... 5 Reallocated_Sector_Ct 0x0033 200 200 140Pre-fail Always - 0 7 Seek_Error_Rate 0x002e 200 200 000Old_age Always - 0 ... 196 Reallocated_Event_Count 0x0032 200 200 000Old_age Always - 0 197 Current_Pending_Sector 0x0032 200 200 000Old_age Always - 0 ... 199 UDMA_CRC_Error_Count0x0032 200 200 000Old_age Always - 0 Cheers, Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/535edd03.5050...@hardwarefreak.com
Re: Personal Recommendations for Free List Compatible Email Service
On 4/25/2014 1:02 AM, Brad Rogers wrote: On Thu, 24 Apr 2014 17:52:13 -0500 Stan Hoeppner s...@hardwarefreak.com wrote: Hello Stan, You're asking specifically about an account for list mail only in this thread. None of these concerns apply. Correct me if I'm wrong (I dare you), but I don't think it's your decision whether the concerns apply to Patrick. I made a simple common sense observation that applies to anyone using a dedicated account for list mail. I didn't think it necessary to explain the blindingly obvious, to quote your sig, but apparently that is necessary after all. So, stating the obvious... One can change a list mail account willy nilly, and the only notification required is changing the subscription address for each list. Thus, again, the concerns Patrick mentioned do not apply to a dedicated list mail account. Given your sniping remark and dare, some predisposition you have against me prevented you from applying the logic of your own signature. Cheers, Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/535a08d0.2080...@hardwarefreak.com
Re: Personal Recommendations for Free List Compatible Email Service
On 4/19/2014 3:38 PM, Patrick Bartek wrote: What with my authentication problems with my Yahoo Mail address on this list, anyone have personal recommendations for a good, free email service ... to run my lists subcriptions through? You have Cox broadband. Why aren't you using Cox IMAP? Most broadband providers offer multiple email accounts per service connection. Create an account for list mail. Done. Cheers, Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/5358d0ae.1030...@hardwarefreak.com
Re: Personal Recommendations for Free List Compatible Email Service
On 4/24/2014 12:03 PM, Patrick Bartek wrote: On Thu, 24 Apr 2014, Stan Hoeppner wrote: On 4/19/2014 3:38 PM, Patrick Bartek wrote: What with my authentication problems with my Yahoo Mail address on this list, anyone have personal recommendations for a good, free email service ... to run my lists subcriptions through? You have Cox broadband. Why aren't you using Cox IMAP? Most broadband providers offer multiple email accounts per service connection. Create an account for list mail. Done. Years ago, when I initially switched to Cox from dialup, and had to deal with the hassle of notifying everyone of my new addresses, I decided I needed email addresses (for business and personal) that wouldn't ever change regardless of where I resided, or who I worked for, or what Internet provider I used, or even especially if I had one. So, that meant Cox was out. Even though I did consider it, I opted to stick with my original decade plus old plan. It's way more practical. You're asking specifically about an account for list mail only in this thread. None of these concerns apply. Cheers, Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/5359959d.6070...@hardwarefreak.com
Re: Heartbleed (was ... Re: My fellow (Debian) Linux users ...)
On 4/13/2014 10:03 PM, Chris Bannister wrote: ... considering it is a catastrophe worse than the Y2K bug. This is several orders of magnitude less severe than Y2K. It seems very likely that people are using compromised apps on their smartphone and you'd think it would be advisable to warn people ASAP! OpenSSL is a library, not an 'app'. Not even an email from the bank! Many/most financial institutions disdain open source software and would much rather pay for proprietary commercial solutions so there is someone to sue and recover damages when things go tits up. Most financial institutions tend to run operations on IBM or clone mainframes. Thus they'll likely be using IBM's mainframe implementations of SSL/TLS, or a commercial front end termination device, neither of which are likely affected by this CVE which is for a few specific version of OpenSSL only. Then there is also the very serious issue of embedded devices using openssl. Tablets, smartphones, routers, ... etc. etc. This problem only exists *if* these devices connect to a compromised or rogue host via SSL/TLS *and* the user hasn't reset and or deleted locally cached usernames and passwords. So, no, definitely not on the impact scale of Y2K. That affected *everyone* whereas this does not. Anyone using an MS Windows PC, which is the majority of the planet, whose financial institutions do not use OpenSSL, are entirely safe from this bug. The *nix community is going ape shit over this not because of bank accounts potentially getting drained, but because so many command/control systems of the Internet backbone are vulnerable to leaking encryption keys, potentially allowing any cracker access to them. Cheers, Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/534b8648.1000...@hardwarefreak.com
Re: Heartbleed (was ... Re: My fellow (Debian) Linux users ...)
On 4/14/2014 5:53 AM, Jochen Spieker wrote: Stan Hoeppner: On 4/13/2014 10:03 PM, Chris Bannister wrote: Then there is also the very serious issue of embedded devices using openssl. Tablets, smartphones, routers, ... etc. etc. This problem only exists *if* these devices connect to a compromised or rogue host via SSL/TLS *and* the user hasn't reset and or deleted locally cached usernames and passwords. That is not the whole truth. Yes, this is the whole truth. It has by now been shown that certificates and private keys were at risk for two years. You are affected by this bug if your browser (or any other SSL/TLS client) does not properly check for certificate revocations or if you try to visit a previously vulnerable system whose certificate was not revoked for some reason. Hence my statement above: connect to a compromised or rogue host So, no, definitely not on the impact scale of Y2K. That affected *everyone* whereas this does not. Anyone using an MS Windows PC, which is the majority of the planet, whose financial institutions do not use OpenSSL, are entirely safe from this bug. No. This applies to everyone who is using sites that previously used a vulnerable version of OpenSSL. Since I generally cannot know which software is used by a specific site, I tend to go as far as concluding that any certificate from before 2014-04-08 may be stolen. Intentionally quoting me out of context and then attempting to correct my factual statements, without adding anything constructive to the thread. That's trolling. BTW, you shouldn't focus only on banks either. There are a lot of popular services that use free software a lot, some of which happen to include payment functionality. I did not focusing on banks. I replied to Chris Bannister's statement regarding *his bank*, which you snipped, again intentionally deleting context in order to be a contradictarian. Might have to add you to the kill file... Cheers, Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/534bc714.4040...@hardwarefreak.com
Re: Heartbleed (was ... Re: My fellow (Debian) Linux users ...)
On 4/14/2014 6:41 AM, Richard Hector wrote: On 14/04/14 23:31, Stan Hoeppner wrote: BTW, you shouldn't focus only on banks either. There are a lot of popular services that use free software a lot, some of which happen to include payment functionality. I did not focusing on banks. I replied to Chris Bannister's statement regarding *his bank*, which you snipped, again intentionally deleting context in order to be a contradictarian. Chris, like me, appears to be in New Zealand. The only local bank I've heard any info about is Kiwibank, who are apparently not vulnerable due to running their systems on Windows. So they're just vulnerable to everything else... I believe at least one local bank runs most of their stuff on Linux, but I haven't heard anything from them. Perhaps (some of the) banks are a bit smaller here, and don't necessarily run to the mainframes used elsewhere. I certainly wouldn't jump to conclusions that they're a bank therefore they use IBM mainframes therefore they don't use OpenSSL therefore they're invulnerable, I jumped to no conclusion. Do you see the word bank in my original statement below? No, you see financial institutions. and I wish that they'd tell us either way. Yes, that would be nice. But outside of technical geeks, none of their customers are paying attention. And, more importantly, as a rule chiseled in granite, financial institutions, especially banks, never admit to doing anything wrong, because it opens them up to liability, lawsuits, thus monetary loss. The lawyers have sewn the executives lips shut on this while they spend days, if not weeks to a month figuring out how to best handle needed disclosure without losing [m|b]illions. On 4/14/2014 1:55 AM, Stan Hoeppner wrote: Many/most financial institutions disdain open source software and would much rather pay for proprietary commercial solutions so there is someone to sue and recover damages when things go tits up. Most financial institutions tend to run operations on IBM or clone mainframes. Thus they'll likely be using IBM's mainframe implementations of SSL/TLS, or a commercial front end termination device, neither of which are likely affected by this CVE which is for a few specific version of OpenSSL only. Financial Institutions, not an exhaustive list: banks credit unions credit/debit card companies - VISA/MasterCard/etc credit/debit card processors - Paymentech, etc exchanges - stock and mercantile, dozens of them worldwide NYSE, NASDAQ, London, Hong Kong, Tokyo, Chicago Merc brokerage houses - hundreds worldwide fund management companies - pensions, mutual funds, IRAs, etc etc, etc Cheers, Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/534beac5.5020...@hardwarefreak.com
Re: Maildir ext4 slowness
On 4/8/2014 9:47 PM, Kumar Appaiah wrote: Dear Debian User, Of late, I've observed that opening my Maildir boxes in Mutt has been a tad slow. Here is the rough structure: I have an LVM home directory (ext4), within which I have a folder called ~/Maildir. This folder has several Maildirs, say inbox, debian-user etc., each of which gets its mail delivered using procmail. tune2fs gives me these features: has_journal ext_attr resize_inode dir_index filetype needs_recovery extent flex_bg sparse_super large_file huge_file uninit_bg dir_nlink extra_isize Of late, I've observed that folders with over 1000 messages seem quite slow to respond (order of 3-4 seconds), which wasn't really happening in the old days when I was using mboxes. I prefer Maildir since I can use notmuch to index my mails easily. Is there something I could to to speed things up? For instance, I could create a 20 GB file, create another filesystem on it and mount it as my Maildir, if that would help. Sparse files tend to fragment horribly, especially on EXT4, and this will be exacerbated by storing maildir files in it. So this is not your solution. If anything it would be worse than now. As with any MUA directly accessing maildir files performance gradually slows down over time with more and more mail files because they are scattered across the filesystem, especially with EXT, much less so with XFS. Seeking to 1000 files all over the disk to read the headers takes time. mbox is quicker because you typically have far more messages stored linearly on disk, thus you have much less seeking when reading headers. Mutt, as with other MUAs, creates a header cache so it only needs to read the headers of new mail files. If you're slowing down with large maildir folders, the most likely problem is that your header caching is not working properly. I am not a mutt user so I cannot tell you which knob to turn or which cache file to delete/recreate to fix this. I have seen the same problem with many other MUAs and this is most often the cause. See this thread from '08 for background on the excessive seeking. The patch for this should already be in your mutt version if it's current, but understanding what is being discussed may give you an idea as to what is broken, and where to start looking: http://lkml.iu.edu//hypermail/linux/kernel/0812.2/00514.html Cheers, Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/5344d260.2010...@hardwarefreak.com
Re: Gigabyte mother board AMD cpu
On 3/26/2014 8:17 AM, Dan Ritter wrote: On Wed, Mar 26, 2014 at 11:19:42PM +1100, Ike Shields wrote: I have Gigabyte mother board with AMD cpu No. GA-78LMT-S2P I need the drivers for it. I am using Debian Release 7.4 (wheezy) 64-bit Everything on that motherboard should work in Debian. The onboard video card needs xserver-xorg-video-radeon or fglrx-driver from non-free. The ethernet wants firmware-realtek. USB3 should work immediately. Anything else? Realtek ALC889 audio chip is supported by alsa. The hardware monitoring chip, iTE IT8720, has been supported by lm-sensors since 2008. This chip also provides the PS/2 and serial port functions, both of which are supported. Cheers, Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/5332fd65.1010...@hardwarefreak.com
Re: Spreading NIC interrupts across multiple CPUs
On 3/26/2014 1:28 PM, Aaron Seelye wrote: I have a question regarding interrupt balancing for a NIC across CPUs. I have a Dell R710 (dual quad core) with embedded broadcom 5709 that seems to put everything on the CPU0. I even threw an Intel Pro/1000 PT in the Dell, but this is showing the same problem. For a test system, I have an HP DL360-G5 (also dual quad core) with embedded broadcom 5708 that balances across all cores. I've also thrown in an identical Intel NIC, and it seems to balance across the cores properly. This leads me to believe that there's something wrong with my BIOS setup, or there's something inherently wrong with the R710, though I'm leading towards the former, as I'm seeing this on two R710s, and doubt I'd hit a magic breakage across two chassis. Also, this is with no massaging on my part, both running up to date debian wheezy 7.4, with the Dell being installed originally with 7.1 My question is this, what option(s) could be present with the R710 bios that would cause something like this to happen? If not the bios, where/what else should I look at? Please read this for educational background, especially the Note at the bottom of the page. https://access.redhat.com/site/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/Performance_Tuning_Guide/s-cpu-irq.html Then ask an intelligent question about IRQ balancing and steering, WRT the two specific and different hardware systems, and Debian kernel versions, being used on each. Cheers, Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/53334a40.8020...@hardwarefreak.com
Re: Spreading NIC interrupts across multiple CPUs
On 3/26/2014 5:23 PM, Aaron Seelye wrote: On 3/26/2014 2:44 PM, Stan Hoeppner wrote: Please read this for educational background, especially the Note at the bottom of the page. https://access.redhat.com/site/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/Performance_Tuning_Guide/s-cpu-irq.html Then ask an intelligent question about IRQ balancing and steering, WRT the two specific and different hardware systems, and Debian kernel versions, being used on each. I'd seen other things similar to that, however, it doesn't seem to get me any closer to the solution. Please post the full output of cat /proc/interrupts without line wrapping. The output from one of the Dell (not balanced) systems: root@conf-2:~# uname -a Linux conf-2 3.2.0-4-amd64 #1 SMP Debian 3.2.54-2 x86_64 GNU/Linux root@conf-2:~# grep eth /proc/interrupts 79: 704642666 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 PCI-MSI-edge eth0 root@conf-2:~# cat /proc/irq/79/smp_affinity root@conf-2:~# cat /proc/irq/79/smp_affinity_list 0-15 This is an 8 core machine with HT enabled, 16 logical CPUs, so right off the bat it is dramatically different than the Compaq machine below as far as the kernel is concerned and how scheduling is performed. The current mask may or may not be correct for this configuration. I never use HT and I can't find any docs about HT and /proc/irq/xx/smp_affinity. If this is a production machine and you can't easily reboot it to disable HT, first try a mask that includes only the physical CPUs and not the logical: ~# echo ff /proc/irq/79/smp_affinity This should schedule IRQs only on the 1st logical processor (physical CPU) of each core. If that doesn't do the trick reboot the box and disable HT. If that doesn't do it I'll dig further into the scheduler to figure out what's going on. The output from the HP (balanced) system: root@deb-test:~# grep eth /proc/interrupts 68: 4251 4190 4212 4264 4226 4257 4251 4214 PCI-MSI-edge eth0 root@deb-test:~# cat /proc/irq/68/smp_affinity ff root@deb-test:~# cat /proc/irq/68/smp_affinity_list 0-7 This is an 8 core machine without HyperThreading. The mask is correct for 8 physical CPUs. Oddly though, one box outputs the leading zeros of the mask while the other does not. Or did you mung either output? As you can see, both systems are running identical kernels, and both have affinity set to spread across all CPUs. The latter may not be a correct statement, as HT logical processors are not CPUs. Also, the smp_affinity mask on the Dell implies 32 processors. Many, but not all, of the functional units are duplicated. Just as you do not want to schedule two compute intensive tasks to both logical processors on a core leaving the other cores idle, you also do not want to assign assign any interrupts to the 2nd logical processor in a given core. All this does is pile up context and state switches on said core. The net effect is decreasing the overall work that can be performed. And to this point, it's not usually a good idea to spread interrupts round robin from any device evenly across all cores in a system. This is inefficient as each core must load the ISR for every interrupt. This decreases the effectiveness of L1/L2 caches on all cores, causing additional cache misses for other processes executing on those cores. This is precisely why irqbalance was created. However, the Dell is using CPU0 exclusively for the ethernet device interrupts, while the HP spreads them pretty evenly. This could be as simple at HT being enabled on the Dell. If not, the contents of your /proc/interrupts files should help me narrow this down for you. For future reference, kernel scheduler problems such as this should be posted on LKML, not a distro list, no matter which distro you use. There are very few people on debian-user or any of the distro general help lists with significant knowledge of the kernel, let alone the scheduler. You typically get help with this kind of thing much faster, and with more thorough knowledge transfer on LKML. Cheers, Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/5333b78c.9090...@hardwarefreak.com
Re: On what is helpful and what is not [was: Re: Wifi]
Note: My use of you/your etc is plural, singling out no individual Note: To everyone. Don your Kevlar and CBRN suits ;) I've been avoiding engaging this juvenile nonsense, but given some of you whiny thin skinned PC police simply will not drop this bone, apparently it's up to me to lay down the law. I'm blunt, I'm curt, I'm brash, I'm confident. I often ruffle the feathers of thin skinned users, mostly new but on occasion old hands. I make no apologies because I get things done for people on this list, and on occasion am the only member here who can get them the information they need, at the level of technical depth required. Constantly worrying about the thickness of others' skin during the process is not my responsibility, *especially* thin skinned whiny children who aren't even the recipient of an incorrectly perceived insult. Mission, project, task oriented people tend be singularly focused on the matter at hand. We don't stand around the water cooler half the day complimenting each other on our pretty new shoes and hairdos to boost each others' self esteem. We practice tough love when needed, and we don't coddle people. We enable them. We may be perceived as abrasive or insulting because we are focused on the mission and not focused on worrying about other people's fragile feelings. Their fragility is their problem, not mine. Some might describe this as akin to a military mindset, where emotions are expressly ignored, trained out of a person. I am apparently the only person on this list who practices tough love, and I'm not one bit shy about it. Posts like the OP's demand tough love, otherwise people will EXPECT to be coddled every time they have a problem. I refuse coddle people. That's giving a heroin addict free heroin. It's a disservice to them because it trains them to NOT learn and perform problem solving on their own. The give a man a fish or teach him to fish adage. It's a disservice to myself because it unnecessarily wastes valuable time, just like the need to spank people here for this childish behavior. If you want to coddle users, that's your choice. I never complain about your style do I? Likewise, if I choose to be blunt, brash, and practice tough love, which benefits myself and the other user, no one has the right to complain about my style. But the fact is, many are complaining, whining like a bunch of children. Mommy! Mommy! The bully at school called another kid a nub! Wa-wa-wa-wa-wa! This is absolutely ridiculous behavior, and it's exactly what you've been engaged in. On 3/9/2014 7:15 PM, Dave Woyciesjes wrote: ... Really, calling the OP a nub ( whatever the hell he means by that) isn't an insult? And you think there was no condescending insulting tone overall? Dave, every aspect of life exists in a context, not a vacuum. Re-read this paragraph and explain how the context of my use of nub is an intention to levy an insult, not simply use of the phonetic shorthand for newb. On 3/8/2014 11:35 PM, Stan Hoeppner wrote: In fact, given he assumes everyone knows why he's punching the function keys, it's pretty certain he's a nub. And that's fine. But we need to know his knowledge level in order to best assist him. Do you notice the sentence And that's fine. directly after nub? And the sentence directly after that? Is this the context of an insult? Clearly not. If I was using nub as an insult, why would I immediately qualify that with And that's fine.? I wouldn't, and neither would anyone else. The context here is clear. I am using nub as a general description of knowledge level, and then stating we need to ascertain where in that range his knowledge level is in order to tailor our responses. The context of nub as an insult is simple: You're a f--king nub! It's hard to miss the difference between the two contexts is it not? So how on earth did you confuse the contexts? Tunnel vision bias? I've been using the phonetic abbreviation nub for over 15 years. Just because some urban dictionary, or some segment of the net populous decided in the last 5 years, 8 years, etc, that this is not equivalent to newb, is categorically an insult, no matter who uses it, or in any context, doesn't not make it so, just as I can declare red and blue are now swapped, which doesn't make it so. It's pretty damn sad that a 'certain' type of new users join this list, and the first post they see from me is one of my rare tough love posts such as in this thread. And even though said new users are not the recipient of the comment, they nonetheless instantly overreact, blow their top, and try to rally the list to tackle the bully. What thoughtful people do is create mail filters, without uttering a peep on or off list. I know I've been extremely vague, apologetic, and politically correct here, keeping my opinions so close to the vest. So I apologize if it has been difficult for you to decipher my cryptic comments. Cheers, -- Stan
Re: Wifi
On 3/9/2014 5:09 AM, Andrei POPESCU wrote: ... Nope, pretty much everyone was stabbing in the dark due to lack of information. Now, there's a good chance one of these stabs actually hits the target (my money is on firmware), but pretty much everyone RTL non-free firmware was the first thing that popped into my head as well, or possibly the NIC wasn't detected during install thus no module loading. That's before I typed my first reply. Just not enough info. Subsequent research showed that Toshiba model could have any of 3 NIC models, Atheros, Intel, or Realtek. The Intel NICs AFAIK are all free firmware, so if he has the Intel NIC it shouldn't be a firmware load issue. I dunno the non/free status of the Atheros firmware, too lazy to look it up ATM. assumed the OP is able to identify the correct firmware package and install it by himself (if his WiFi is not working he might not even have easy access to the internet). I'd say he definitely has another box handy, which he posted his original message from. Guessing it's not Debian, given his first post comments, and that it was sent from Thunderbird instead of IceDove. though the latter point is certainly not definitive: User-Agent: Mozilla/5.0 (X11; Linux i686; rv:24.0) Gecko/20100101 Thunderbird/24.3.0 Now that I've discovered his other system (at least one, maybe more) is Linux, I'd guess he's not a total... novice. He may have simply been in a big hurry when he posted and didn't consider what relevant info was needed. Surely we'll know before long. I will offer the following. Regardless of the posting styles of the various personalities on this least, this OP, and others, should be grateful for the depths to which folks here will dig to get an issue resolved, no matter how sparse the problem report, how little to go on. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/531c48c9.6080...@hardwarefreak.com
Re: Wifi
On 3/9/2014 6:36 AM, Andrei POPESCU wrote: On Du, 09 mar 14, 05:56:09, Stan Hoeppner wrote: The Intel NICs AFAIK are all free firmware, so if he has the Intel NIC it shouldn't be a firmware load issue. Unfortunately not :( $ dmesg | grep firmware [9.914262] iwl4965 :03:00.0: firmware: direct-loading firmware iwlwifi-4965-2.ucode [9.914276] iwl4965 :03:00.0: loaded firmware version 228.61.2.24 Yeah, I'd forgotten how stupid, yes stupid, the Debian policy on source code is, free distribution of binary firmware blobs be dammed. The only thing this accomplishes is pissing off users, especially new ones. The targets of the policy, the hardware vendors, haven't budged one micrometer in all these years because Debian swings a twig, not a heavy club such as Red Hat or SuSE. Debian will never posses a large enough club, so this policy should be abandoned for the sake of the (especially new) users. I started rolling my own kernels from vanilla source 10 years ago, and from that point I've built my drivers and firmware blobs into the kernel. So I've not had to deal with Stallman's rabid socialist nonsense for a very long time. I solved this problem long ago and moved on. Out of sight, out of mind as they say. I actually thought Atheros chipsets didn't need firmware, but the description of firmware-atheros seems to suggest otherwise. Atheros have had many generations of wireless product. The vanilla 3.2 source tree has drivers for 4 generations of Atheros NICs, 3.13 has 8. ../carl9170/fc.c performs full byte verification of the eeprom data at driver initialization, so it seems clear this model requires host (kernel) installed firmware. I don't find a similar file in the other ath directories, suggesting the other models may have non-installable, or factory loaded permanent firmware. I say 'may' because I simply don't have time to browse the remarks and code in all of these many dozens of wireless driver files. I think it's safe to say that some Ath cards have used installable firmware and some don't. The requirement of loadable firmware has nothing to do with Linux, but the desire of the hardware vendors to be able to fix problems and add functionality over the life of the product, primarily the former. Most of it is designed for Windows, and vendor products that have passed WHQL certification can get new drivers and firmware pushed out via Windows Update. Installation is optional, but the automatic distribution infrastructure is there. Cheers, -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/531cb100.8040...@hardwarefreak.com
putting the 'wifi switch' argument to rest
Nobody is 'right'. We're all wrong to some degree. I based my assertion that there is no such thing as a WiFi Switch on the fact that the term/phrase has been co-opted by lay and marketing people to describe all manner of things related to Wifi--not a singular thing--including, but not limited to: WiFi controlled light switch WiFi remotely switched AC power outlet Wireless Router with wired Ethernet switch Laptop WiFi antenna [en|dis]able via physical or software switch As it turns out, there is an actual device whose official name is WiFi Switch. It's a solid state integrated circuit switch, actually a family of IC switches, present in every wireless device, whose functions are one or more of the following, depending on IC model, in order of functional capability importance: 1. Switch at high frequency between transmit and receive ports on the network interface IC because radio is not a full duplex technology 2. Switch amongst multiple antennae on devices such as routers and APs possessing 2 or more antennae, simultaneously with #1 in some designs 3. Switch between different network device IC ports, such as a cell network IC and 802.11 IC, or simply different ports on once IC if it integrates both physical protocols--used in smart phones, etc Some, but not all, of these IC also have a no-connect switch position, which provides the facility for disabling the antenna(e). This one function of many, not included on all switch ICs, and the least important of the bunch, is apparently what most lay people consider to be the sole function of a WiFi Switch. I always take exception to lay person use of terminology, with good reason, because the meaning can be different in the minds of different people. That's the purpose of standards. In this thread, everyone but me has been using one of many layperson definitions of WiFi Switch, not the technically correct definition presented above. The following document from a leading manufacturer of WiFi Switch ICs provides the technical data proving my points above: http://www.rfmd.com/CS/Documents/BR_Switch_LNA_FE_Solutions.pdf Maybe this can finally put this time wasting sub thread to rest. Cheers, -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/531d3c80.3060...@hardwarefreak.com
Re: Wifi
On 3/8/2014 2:18 PM, Patrick Alouidor wrote: Hello all. I'm not sure if it me but I have a fresh install of Debian 7 on laptop Toshiba C-55A5310. and For some reason I cannot enable my wifi switch. I have been pressing the F keys but no luck. please This is my first Laptop ever and I wanted to put something stable on it and now I cannot get my wifi to turn on. My I please get some form of assistance on wifi. This is an invalid and incomplete help request. You must perform basic troubleshooting and tell us specifically what is not working, or we can't really help you. You mention a wifi switch. There is no such thing. The laptop has a wireless ethernet adapter usually of the 802.11 a/b/g/n standard. It will connect to a wireless router or wireless access point. Based on your generic problem report, the issue could be with your router/AP or it could be with the wireless adapter configuration in Debian 7. The lack of detail in your problem report suggests you have never used Linux, or that you've never participated in a technical forum. Please tell us your Linux skill level so we can reply with appropriate level of instruction. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/531be558.50...@hardwarefreak.com
Re: Wifi
On 3/8/2014 10:02 PM, Tom Furie wrote: On Sat, Mar 08, 2014 at 09:51:52PM -0600, Stan Hoeppner wrote: On 3/8/2014 2:18 PM, Patrick Alouidor wrote: Hello all. I'm not sure if it me but I have a fresh install of Debian 7 on laptop Toshiba C-55A5310. and For some reason I cannot enable my wifi switch. I have been pressing the F keys but no luck. please This is my first Laptop ever and I wanted to put something stable on it and now I cannot get my wifi to turn on. My I please get some form of assistance on wifi. You mention a wifi switch. There is no such thing. The laptop has a wireless ethernet adapter usually of the 802.11 a/b/g/n standard. It will connect to a wireless router or wireless access point. Given the context I would surmise that wifi switch means a switch on the laptop to enable/disable the wireless adapter, whether that be an actual switch, button, or key-combo. I would surmise his wifi switch is his wifi enabled cable/DSL router that also has an inbuilt 4 port fast Ethernet or GbE switch, stated in big bold letters on the box, prompting him to call it a wifi switch. Is your guess right or mine? He didn't mention WEP/WAP key setup or any other manual configuration steps/issues, which leads me to, again, guess, that he's trying to do WiFi Protected Setup (WPS) auto configuration. So maybe by wifi switch he means the WPS button on the WiFi router. And maybe Network Manager/WICD use the function keys to initiate WPS auto negotiation. I never do auto anything so again this is a guess. And I'd guess based on his post that WPS is exactly what he's attempting. The whole point of my post was to eliminate the guessing and get right to helping the guy at the technical level, or lack thereof, which he requires. It's pretty clear from his lack of correct terminology and technical details, no initial troubleshooting performed by him, that he's a total nub. All of the replies to this point, but mine, assume he knows how to get a bash shell to run commands and perform other common tasks. He may not even know that much. In fact, given he assumes everyone knows why he's punching the function keys, it's pretty certain he's a nub. And that's fine. But we need to know his knowledge level in order to best assist him. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/531bfd9e.50...@hardwarefreak.com
Re: Wifi
On 3/8/2014 10:20 PM, Ralf Mardorf wrote: On Sun, 2014-03-09 at 04:02 +, Tom Furie wrote: On Sat, Mar 08, 2014 at 09:51:52PM -0600, Stan Hoeppner wrote: On 3/8/2014 2:18 PM, Patrick Alouidor wrote: Hello all. I'm not sure if it me but I have a fresh install of Debian 7 on laptop Toshiba C-55A5310. and For some reason I cannot enable my wifi switch. I have been pressing the F keys but no luck. please This is my first Laptop ever and I wanted to put something stable on it and now I cannot get my wifi to turn on. My I please get some form of assistance on wifi. You mention a wifi switch. There is no such thing. The laptop has a wireless ethernet adapter usually of the 802.11 a/b/g/n standard. It will connect to a wireless router or wireless access point. Given the context I would surmise that wifi switch means a switch on the laptop to enable/disable the wireless adapter, whether that be an actual switch, button, or key-combo. JFTR When I searched the Internet for Toshiba C-55A5310, I didn't found the information what chip is used. http://www.toshiba.com/us/computers/laptops/satellite/C50/C55-A5310 Toshiba has produced the C55-A5310 using 3 different OEM supplied 802.11 cards: Atheros, Intel, and Realtek. This is the same case across their line. In any given 3/6/12 month period they change their wireless NIC supplier to get best price, same with SATA HDD, media drive, SO-DIMM, etc. Anything socketed and standardized. I find on page 156 of the users guide that Fn+F12 enables/disables the wireless antenna. http://cdgenp01.csd.toshiba.com/content/support/manuals/userguides/su4001258/GMAD00349010_13Apr24.pdf So this explains why the OP was hitting the function keys. This antenna function may/not be controlled by an OS driver, so it may/not work with Linux without the appropriate Toshiba driver, if one exists. However, whatever Linux driver his 802.11 NIC uses should have the ability to turn the antenna on/off via command line or network manager. But until we get more info from the OP there's nothing more we can do but guess. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/531c1018.4080...@hardwarefreak.com
Re: whitel...@lists.debian.org
On 3/5/2014 1:08 AM, Ralf Mardorf wrote: Mails from ralf.mardorf-at-alice-dsl.net seldom come through the list. I show 623 from that address since 10/13. What you meant to say is on occasion my mails do not post to the list. For those mails that don't come through, I never got a postmaster reply, IOW no delayed, blackhole listed etc. replies. Then you need to contact the Debian list postmaster and the postmaster of alice-dsl.net, to figure out where the problem is. Telling your story on debian-user won't have any impact whatsoever. My MUA never informed me about issues when sending a mail. I now will replace Evoultion's SMPT with msmpt, perhaps I then will get messages. That's not the problem. All versions of SMTP handle NDRs. If you're not receiving an NDR for messages that disappear then no NDR is being generated. I subscribed or tried to subscribed to the whitelist with my ralf.mardorf-at-alice-dsl.net account, but nothing happened. Why would it? You already have that addressed subscribed to debian-user. The whitelist is for addresses that are not subscribed to any lists. It seems you have a mail problem but you do not know what it is. Instead of troubleshooting to identify the problem, you're attempting to circumvent it with the whitelist nonsense. To troubleshoot this take the following steps: 1. For emails you believe are lost, check the list archive. Are they in the archive? If so they aren't being rejected. In this case, either the MX for your domain is rejecting list messages, or your anti-spam or anti-virus software is eating them. 2. If they are not in the archives, contact the debian.org postmaster and give him the message-id of the lost email. He can tell you if if was received by the Debian mail servers and if they message was posted to the list. 3. If debian.org did not receive the email, then you need to contact the postmaster at alice-dsl.net, provide the message-id, and as if the email was relayed. Standard troubleshooting stuff. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/5316d829.2070...@hardwarefreak.com
Re: fancontrol wheezy Dell T7610
On 2/25/2014 3:47 AM, Dan wrote: I recently bought a workstation to do calculations. It has two Xeon processors with 16 cores and 32 threads in total. I realized that the temperature gets very high on high load typically 80C. That is way too much. Then I changed the fan speed in the bios from auto to high. Now temperatures are reasonable 45C, but it is very noise and it never stops (even with no load) I have to reboot the computer to change the fan speed.. ... Any idea or suggestion? Is it the CPU fans or the chassis fans generating the intolerable noise? If just the CPU fans, you can simply replace them with quieter models. If your system accepts wide coolers, two of this model would be suitable http://www.newegg.com/Product/Product.aspx?Item=N82E16835114120 If it accepts only narrow coolers, then two of these http://www.newegg.com/Product/Product.aspx?Item=N82E16835114142 The 8 core Sandy/Ivy bridge Xeons range from 95-150 watts Thermal Design Power, or 180-300 watts combined for a 2 socket system. The stock style fans required to evacuate this amount of heat, plus that of the GPU, HDDs, system chipsets, DRAM, VRMs, etc are not going to be quiet. Dual socket/professional workstations are generally not very quiet machines. Beyond replacing the CPU coolers there are a number of ways to reduce the noise while still achieving the required airflow for proper cooling. Run the fans at full RPM all the time, while damping the interior of the chassis using acoustic damping pads such as http://www.acoustiproducts.com/en/acoustipack.asp When applied correctly, thoroughly, to all interior surfaces, this will absorb much of the high frequency fan noise emitted by the CPU and chassis fans. I don't use such self stick thin foam pads as they are over priced and the performance isn't that great for the money. I use 1.5 acoustic egg crate foam attached with 3M Super 77, as the egg crate foam yields superior acoustical damping performance. http://www.parts-express.com/acoustic-sound-damping-foam-1-1-2-x-24-x-18-ul-94--260-516 This is the material used inside of speaker cabinets and recording studios. The self stick thin foam is much easier for most people to install which is why I mentioned it first. You can also use a manual fan speed controller such as http://www.newegg.com/Product/Product.aspx?Item=N82E16811999171 This allows you to fine tune noise level vs cooling performance using continuously variable knobs. This requires replacing any existing PWM fans in system with non-PWM fans if using a standard fan controller. There are some PWM controllers on the market but I'd avoid them. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/530cdc9b.7090...@hardwarefreak.com
Re: fancontrol wheezy Dell T7610
On 2/25/2014 9:53 AM, Dan wrote: ... I didn't open the computer. I do not know if the fan connectors have 4 pin. I prefer not to open the computer. It is on warranty. If you don't want to void the warranty, then don't monkey with the fan speed or accidentally shut any fans down for any amount of time. That can roast things. There are companies that sell sound killing computer cabinets for tower style computers but they're not cheap. Google is your friend. You don't strike me as the handy type but I'll suggest the most common inexpensive DIY path taken for killing PC noise. This assumes the T7610 sits on the floor next to your desk, not on it. If not, put it on the floor already. Build a 3 sided box out of 3/4 MDF with outside dimensions, matched to the T7510 11W x 18H x 28D Cover the inside and outside with low pile carpet, using 3M Super77 adhesive or simply lots of T50 1/4 staples. Wrap it around the edges and neatly trim it so it looks half way decent. Lower it over the T7610 so you have 3.5 overhang front and rear. This will pretty much kill the noise problem instantly. Carpeting the outside isn't necessary for killing noise, but looks better than bare MDF. You could use black synth wood grain Melamine or even Maple veneer if want it to really look fancy. Finish is your choice. You could substitute 1.5 thick egg crate acoustic damping foam on the inside. Carpet remnants are usually easier for most people to acquire from a store or relative/friend, probably cheaper as well, and just as effective at killing the noise. I've build a few of these in the past and they just work. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/530d0815.1050...@hardwarefreak.com
Re: fancontrol wheezy Dell T7610
On 2/25/2014 3:28 PM, Jerry Stuckle wrote: ... Encasing the tower in a sound Damping not dampening. To dampen something is to add moisture to it. To damp an object is to lower its resonance frequency. One cannot add moisture to sound waves thus there is no such thing as sound dampening. Please use the correct terminology. Saying sound dampening is like fingernails on a chalk board to audio engineers. On 2/25/2014 4:16 PM, Stan Hoeppner wrote: Build a 3 sided box out of 3/4 MDF... Just ensure there is enough ventilation (remember air has to get in also, not just out) and that you can access the power button, DVD, USB ports, etc. A 3 sided box is by definition open at the front, rear, and bottom, so front panel access is not an issue. Making access holes for these without compromising the sound dampening is probably the hardest part. Not at all. The goal here is not to make the workstation completely silent, but to decrease the SPL of the mid and high frequencies to little more than room background noise level at the ears when seated in the desk chair. An acoustically damped 3 sided box with small front/rear overhangs accomplishes this, in the two ways that matter: 1. The damping material, whether carpet or acoustical egg crate foam, absorbs most of the mid and high frequency sound energy generated by the fans. These sound waves normally radiate not only out the front/rear case vents, but also through the thin sheet steel and plastic panels which tend to resonate at or near these frequencies. In the stock configuration the fan noise radiates in all directions, but not uniformly. 2. Because the sound pressure level of mid/high frequencies drops at a much higher rate off axis from ear position, any sound energy at these frequencies not absorbed by the damping material propagates at floor level out the front and back only. By absorbing most, then directionally focusing the remaining mid/high frequency waves, which are now of greatly decreased amplitude due to the damping material in the overhangs, the noise is barely audible while seated in the chair. You must kneel down to floor level to really hear the fans now. This solution works without compromising access to the machine, or compromising cooling capacity. The T7610 is a true business workstation, with front-to-back only airflow. This 3 sided damping shroud will not work with PCs which have side air intakes, top exhausts, etc. This should be common sense to everyone, but not everyone has common sense, so I'm attempting to head off further me too posts. I tend to only reply to thread topics of which I am a subject matter expert. You made the mistake of assuming that a DIY suggestion implies amateur knowledge, then proceeded to display your truly amateur understanding of the subject matter. I don't post to debian-user that often, but there are folks on this list who know they can take the information and analysis I present straight to the bank. They know the level of expertise and analysis that goes into each and every one of my posts, even those in which I don't give the 2-3 page explanation up front, but the short version which assumes the reader knows a little bit about the subject. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/530d8bf3.1080...@hardwarefreak.com
Re: more than 12G of RAM
On 2/10/2014 10:40 PM, Stan Hoeppner wrote: ... 1. Contact Gigabyte support 2. Buy another identical 8GB DIMM, or exchange this one for two 4s Should have mentioned this sooner. Gary have you flashed the BIOS to the latest rev? As I stated previously, if POST reports 16GB but the e820 map presented to the kernel shows 12GB, this is a BIOS issue. A later BIOS rev may have addressed this, specifically F5. There have been 3 versions for this board, F3/F4/F5. http://www.gigabyte.us/products/product-page.aspx?pid=4642#bios The description for F5: Modified voltage option of NB DRAM compatibility Release date: 08/06/2013 F3 was released in May, F4 in July, F5 in August. The rapid succession of releases, and none since, leads me to believe that most boards probably hit the shelves with the F3 BIOS, and most/all issues were resolved with F5, after most/all board left the factory. If you haven't already, flash the BIOS to F5. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52fab958.3000...@hardwarefreak.com
Re: more than 12G of RAM
On 2/10/2014 2:14 PM, Gary Dale wrote: On 10/02/14 11:24 AM, Dave Woyciesjes wrote: On 02/09/2014 10:14 AM, Jerry Stuckle wrote: On 2/9/2014 8:27 AM, Gary Dale wrote: On 09/02/14 06:50 AM, Stan Hoeppner wrote: On 2/9/2014 3:32 AM, Efraim Flashner wrote: On Sat, 08 Feb 2014 10:20:49 -0500 Gary Dale garyd...@torfree.net wrote: I'm running Jessie on an AMD64 Gigabyte 970A-D3P board with an FX6100 processor. I had 2x4G DDR3 sticks in it but some of the programs I use were causing excessive thrashing. I added a 1x8G DDR3 stick (got a good price on it, much cheaper than adding 2x4G) which resolved the thrashing problem. The BIOS shows I have 16G but free shows only 12G. I also ran free on a machine with 2x8G DDR3 running Wheezy and it showed 16G when I ran free. This suggests that the kernel is handling 16G (as one would expect) in the general case and the issue is likely due to my setup. Since the BIOS shows the full 16G, the problem doesn't seem to be on the mainboard. Is there an issue with running interleaved and non-interleaved RAM together on the Jessie kernels? It sounds to me like you have some issues between your two sets of ram, the 8G stick and the 2x4G sticks. Is there a difference in timings/speed/voltage? I've never put much stock in people saying that you shouldn't mix different types of ram if the price is right, but you might need to change around the placement order. Assuming the motherboard supports dual-channel ram, I'd make sure you have the 2x4G sticks paired up and the 8G stick on its own channel. According to page 16 of the manual you have an unsupported memory configuration: http://download.gigabyte.us/FileList/Manual/mb_manual_ga-970a-d3p_e.pdf If this combo will ever work, the first step is to verify the 8GB stick is in one channel and the two 4GB sticks in the other. If you still don't see all 16GB then disable rank interleaving. If that doesn't fix it, disable channel interleaving. If that doesn't fix it, you may be of luck, and will need to either swap the 8GB stick for a pair of matched 4GB sticks, or acquire another identical 8GB stick. That page just tells you how to install dual-channel DDR3 sticks. Again, the BIOS detects the full 16G. This shows that the setup does work with the board. Detected and Working memory are two entirely different things. It is perfectly possible for your BIOS to detect 16GB, but 4GB of it not work properly. Both Stan and Efraim have good comments. I suggest you follow them. Jerry Have you run memtest yet? Yes. After the other comments, I gave it a try. It also only sees 12G. I find this confusing in that the board apparently works with a mixture of interleaved and non-interleaved memory - as witnessed by the fact that I'm running that way. Why should it see the non-interleaved module as only 4G instead of 8G? The board has 4 sockets and according to the specifications supports up to 32G of memory, which means it should support 8G modules. I could understand the problem if it only supported 4G modules, then it might only see part of the extra memory. I could understand it also if the configuration simply didn't work and only showed me 8G total. 12G I don't understand. Apparently you (and everyone else) missed my last post wherein I explained what the problem is here. Here is the explanation a 2nd time: POST displays a message on the screen that 16GB is present. But that subroutine is separate from the BIOS code that generates the e820 memory map that is presented to the kernel, which can be found at the start of your dmesg log, e.g. BIOS-provided physical RAM map: BIOS-e820: - 0009f000 (usable) BIOS-e820: 0009f000 - 000a (reserved) BIOS-e820: 000f - 0010 (reserved) BIOS-e820: 0010 - bddde000 (usable) BIOS-e820: bddde000 - bde0e000 (ACPI data) BIOS-e820: bde0e000 - d000 (reserved) BIOS-e820: fec0 - fee1 (reserved) BIOS-e820: ff80 - 0001 (reserved) BIOS-e820: 0001 - 00083efff000 (usable) POST says one thing, the e820 map another. Which is why neither Debian nor memtest (which uses the Linux kernel) can find the other 4GB. The memory map being presented to Linux is wrong. If the board does not present the correct e820 map in non-interleaved mode witn mismatched stick sizes, then I guess you could call this a BIOS bug. It is possible to work around this by manually creating a proper map. This can be done using kernel command line options. However, for a non kernel hacker this job will require far more time, research, heartache, etc, than the cost of simply swapping modules to allow the board BIOS to work. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: more than 12G of RAM
On 2/10/2014 4:28 PM, Gary Dale wrote: On 10/02/14 03:48 PM, Stan Hoeppner wrote: On 2/10/2014 2:14 PM, Gary Dale wrote: On 10/02/14 11:24 AM, Dave Woyciesjes wrote: On 02/09/2014 10:14 AM, Jerry Stuckle wrote: On 2/9/2014 8:27 AM, Gary Dale wrote: On 09/02/14 06:50 AM, Stan Hoeppner wrote: On 2/9/2014 3:32 AM, Efraim Flashner wrote: On Sat, 08 Feb 2014 10:20:49 -0500 Gary Dalegaryd...@torfree.net wrote: I'm running Jessie on an AMD64 Gigabyte 970A-D3P board with an FX6100 processor. I had 2x4G DDR3 sticks in it but some of the programs I use were causing excessive thrashing. I added a 1x8G DDR3 stick (got a good price on it, much cheaper than adding 2x4G) which resolved the thrashing problem. The BIOS shows I have 16G but free shows only 12G. I also ran free on a machine with 2x8G DDR3 running Wheezy and it showed 16G when I ran free. This suggests that the kernel is handling 16G (as one would expect) in the general case and the issue is likely due to my setup. Since the BIOS shows the full 16G, the problem doesn't seem to be on the mainboard. Is there an issue with running interleaved and non-interleaved RAM together on the Jessie kernels? It sounds to me like you have some issues between your two sets of ram, the 8G stick and the 2x4G sticks. Is there a difference in timings/speed/voltage? I've never put much stock in people saying that you shouldn't mix different types of ram if the price is right, but you might need to change around the placement order. Assuming the motherboard supports dual-channel ram, I'd make sure you have the 2x4G sticks paired up and the 8G stick on its own channel. According to page 16 of the manual you have an unsupported memory configuration: http://download.gigabyte.us/FileList/Manual/mb_manual_ga-970a-d3p_e.pdf If this combo will ever work, the first step is to verify the 8GB stick is in one channel and the two 4GB sticks in the other. If you still don't see all 16GB then disable rank interleaving. If that doesn't fix it, disable channel interleaving. If that doesn't fix it, you may be of luck, and will need to either swap the 8GB stick for a pair of matched 4GB sticks, or acquire another identical 8GB stick. That page just tells you how to install dual-channel DDR3 sticks. Again, the BIOS detects the full 16G. This shows that the setup does work with the board. Detected and Working memory are two entirely different things. It is perfectly possible for your BIOS to detect 16GB, but 4GB of it not work properly. Both Stan and Efraim have good comments. I suggest you follow them. Jerry Have you run memtest yet? Yes. After the other comments, I gave it a try. It also only sees 12G. I find this confusing in that the board apparently works with a mixture of interleaved and non-interleaved memory - as witnessed by the fact that I'm running that way. Why should it see the non-interleaved module as only 4G instead of 8G? The board has 4 sockets and according to the specifications supports up to 32G of memory, which means it should support 8G modules. I could understand the problem if it only supported 4G modules, then it might only see part of the extra memory. I could understand it also if the configuration simply didn't work and only showed me 8G total. 12G I don't understand. Apparently you (and everyone else) missed my last post wherein I explained what the problem is here. Here is the explanation a 2nd time: POST displays a message on the screen that 16GB is present. But that subroutine is separate from the BIOS code that generates the e820 memory map that is presented to the kernel, which can be found at the start of your dmesg log, e.g. BIOS-provided physical RAM map: BIOS-e820: - 0009f000 (usable) BIOS-e820: 0009f000 - 000a (reserved) BIOS-e820: 000f - 0010 (reserved) BIOS-e820: 0010 - bddde000 (usable) BIOS-e820: bddde000 - bde0e000 (ACPI data) BIOS-e820: bde0e000 - d000 (reserved) BIOS-e820: fec0 - fee1 (reserved) BIOS-e820: ff80 - 0001 (reserved) BIOS-e820: 0001 - 00083efff000 (usable) POST says one thing, the e820 map another. Which is why neither Debian nor memtest (which uses the Linux kernel) can find the other 4GB. The memory map being presented to Linux is wrong. If the board does not present the correct e820 map in non-interleaved mode witn mismatched stick sizes, then I guess you could call this a BIOS bug. It is possible to work around this by manually creating a proper map. This can be done using kernel command line options. However, for a non kernel hacker this job will require far more time, research, heartache, etc, than the cost of simply swapping modules to allow the board BIOS to work. Here's what I get in dmesg re
Re: more than 12G of RAM
On 2/9/2014 3:32 AM, Efraim Flashner wrote: On Sat, 08 Feb 2014 10:20:49 -0500 Gary Dale garyd...@torfree.net wrote: I'm running Jessie on an AMD64 Gigabyte 970A-D3P board with an FX6100 processor. I had 2x4G DDR3 sticks in it but some of the programs I use were causing excessive thrashing. I added a 1x8G DDR3 stick (got a good price on it, much cheaper than adding 2x4G) which resolved the thrashing problem. The BIOS shows I have 16G but free shows only 12G. I also ran free on a machine with 2x8G DDR3 running Wheezy and it showed 16G when I ran free. This suggests that the kernel is handling 16G (as one would expect) in the general case and the issue is likely due to my setup. Since the BIOS shows the full 16G, the problem doesn't seem to be on the mainboard. Is there an issue with running interleaved and non-interleaved RAM together on the Jessie kernels? It sounds to me like you have some issues between your two sets of ram, the 8G stick and the 2x4G sticks. Is there a difference in timings/speed/voltage? I've never put much stock in people saying that you shouldn't mix different types of ram if the price is right, but you might need to change around the placement order. Assuming the motherboard supports dual-channel ram, I'd make sure you have the 2x4G sticks paired up and the 8G stick on its own channel. According to page 16 of the manual you have an unsupported memory configuration: http://download.gigabyte.us/FileList/Manual/mb_manual_ga-970a-d3p_e.pdf If this combo will ever work, the first step is to verify the 8GB stick is in one channel and the two 4GB sticks in the other. If you still don't see all 16GB then disable rank interleaving. If that doesn't fix it, disable channel interleaving. If that doesn't fix it, you may be of luck, and will need to either swap the 8GB stick for a pair of matched 4GB sticks, or acquire another identical 8GB stick. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52f76b96.9030...@hardwarefreak.com
Re: more than 12G of RAM
On 2/9/2014 7:27 AM, Gary Dale wrote: On 09/02/14 06:50 AM, Stan Hoeppner wrote: On 2/9/2014 3:32 AM, Efraim Flashner wrote: On Sat, 08 Feb 2014 10:20:49 -0500 Gary Dale garyd...@torfree.net wrote: I'm running Jessie on an AMD64 Gigabyte 970A-D3P board with an FX6100 processor. I had 2x4G DDR3 sticks in it but some of the programs I use were causing excessive thrashing. I added a 1x8G DDR3 stick (got a good price on it, much cheaper than adding 2x4G) which resolved the thrashing problem. The BIOS shows I have 16G but free shows only 12G. I also ran free on a machine with 2x8G DDR3 running Wheezy and it showed 16G when I ran free. This suggests that the kernel is handling 16G (as one would expect) in the general case and the issue is likely due to my setup. Since the BIOS shows the full 16G, the problem doesn't seem to be on the mainboard. Is there an issue with running interleaved and non-interleaved RAM together on the Jessie kernels? It sounds to me like you have some issues between your two sets of ram, the 8G stick and the 2x4G sticks. Is there a difference in timings/speed/voltage? I've never put much stock in people saying that you shouldn't mix different types of ram if the price is right, but you might need to change around the placement order. Assuming the motherboard supports dual-channel ram, I'd make sure you have the 2x4G sticks paired up and the 8G stick on its own channel. According to page 16 of the manual you have an unsupported memory configuration: http://download.gigabyte.us/FileList/Manual/mb_manual_ga-970a-d3p_e.pdf If this combo will ever work, the first step is to verify the 8GB stick is in one channel and the two 4GB sticks in the other. If you still don't see all 16GB then disable rank interleaving. If that doesn't fix it, disable channel interleaving. If that doesn't fix it, you may be of luck, and will need to either swap the 8GB stick for a pair of matched 4GB sticks, or acquire another identical 8GB stick. The devil is always in the details. That page just tells you how to install dual-channel DDR3 sticks. Again, It tells you exactly how to install the sticks, and that each pair needs to match. the BIOS detects the full 16G. This shows that the setup does work with the board. POST displays a message on the screen that 16GB is present. But that subroutine is separate from the BIOS code that generates the e820 memory map that is presented to the kernel, e.g. BIOS-provided physical RAM map: BIOS-e820: - 0009f000 (usable) BIOS-e820: 0009f000 - 000a (reserved) BIOS-e820: 000f - 0010 (reserved) BIOS-e820: 0010 - bddde000 (usable) BIOS-e820: bddde000 - bde0e000 (ACPI data) BIOS-e820: bde0e000 - d000 (reserved) BIOS-e820: fec0 - fee1 (reserved) BIOS-e820: ff80 - 0001 (reserved) BIOS-e820: 0001 - 00083efff000 (usable) This is exactly what is happening in the OP's case--POST says one thing, e820 another. It is possible to manually create a proper map using kernel command line options so the kernel sees all the memory. However, for a non kernel hacker this job will require far more time, research, etc, than the cost of swapping modules. Thus, if moving the DIMMs around doesn't allow the BIOS to create the proper e820 map, the OP's best option is to buy another matching 8GB stick, or swap it for two matching 4GB sticks. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52f80b81.5020...@hardwarefreak.com
Re: How to setup a simple email server?
On 1/25/2014 1:09 PM, Garry wrote: I would like to setup a simple email server and run it out of my house. I have everything needed in order to do it. In fact I had one setup successfully about a year ago and crashed it. I can't figure out how I did it. There's only two email addresses I would like to setup. I would like to use postfix and dovecot; I don't need MySQL. My ISP seems to block port 587; all other ports are open (25, 110, etc.) I have issues wrapping my mind around setting up mx records. The domain is registered on namecheap.com My IP address is static ipv4. I'm running (would like to keep running) Debian 6 32bit. I've followed all the various guides that pull up in search; with each I run into problems I can't seem to resolve. Can someone point me to a tutorial or provide me with some resources I can follow? I am very appreciative with any help willing to be offered. Thank you. Not everyone is fighter pilot or race car driver material. Likewise not everyone is mailop material. If you've made multiple attempts to set this up and failed each time, I'd say it's pretty clear you simply lack the aptitude and prerequisite IP networking and DNS knowledge. How on earth will you even begin to troubleshoot when something breaks, when you have no understanding of how the various layers of the mail stack actually work, how SMTP works? When your mail stops arriving the first thing you will consider is that there a problem with Postfix or Dovecot, that -you- did something wrong. You'll blame yourself because you know you are the weak link. So you'll start monkeying with your Postfix configuration and break it, even though its working correctly. In reality, 99% of the time mail flow problems are due to a routing problem within your broadband provider's network, a down or flapping link between your provider and one of its peers or somewhere else on the net, a DNS lookup issue possibly caused by the former network issues, etc, etc. Running an MX MTA/mailbox server is 90% networking, 10% MTA/mailbox management, especially for a family mail server with 2 mailboxes connected via less than reliable broadband, no SLA, etc. The first tools that get pulled out for diagnosing an MX problem are dig, traceroute, ping, telnet, etc. NOT 'postfix reload'. Not 'vi /etc/postfix/main.cf' or 'vi /etc/postfix/master.cf'. If you don't have sufficient experience with such network diagnostic tools, don't know how to use them, why you're using them, or when to use which one, you don't have any business running an MX MTA. Attempting to do so will simply cause you problems and burn huge amounts of your time, and will completely overshadow any benefits you might receive from having your mailboxes in a server on the premises. Setting up Postfix and Dovecot should take no more than a couple of hours each, most of it reading docs. Creating your MX and SPF records takes a few minutes, though it may be many hours before they go live depending on your SOA TTL. Programming an inbound TCP 25 PAT mapping on your router should take only a few minutes. None of this is particularly difficult. Figuring out what ports your provider does/not block may be more difficult, as well as getting those blocks removed. If you have a static IP, and especially if you pay extra for it, this shouldn't be a problem. I am not indicting your aptitude nor abilities. I am simply giving you a healthy dose of reality so your eyes are wide open going forward. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52e4d30b.3010...@hardwarefreak.com
Re: Install Debian on SD
On 1/22/2014 11:22 AM, Bob Proulx wrote: Another thing that I did was to install 'eatmydata' in order to use the LD_PRELOAD library to disable fsync(). A lot of applications have added fsync() calls everywhere to disable the file system buffer cache. fsync() doesn't disable the buffer cache but bypasses it for a given write. I wanted to minimize writes to the SD card. Applications that use fsync() are forcing writes through the cache to storage. This To make sure data is on disk before power fails, or the system crashes, etc. fsync() is a good thing. does two bad things. Defeating the cache is many times slower. Quality, mature filesystems use fsync() only for journal and metadata writes, not data, to keep the filesystem in a consistent state after mishaps. Thus speed is application or workload dependent. An MTA workload using maildir mailbox format will take a hit as it is metadata heavy. A MythTV server won't notice the difference as there is nearly zero metadata in the workload. Defeating the cache causes a lot of I/O for flash storage. This is patently false. fsync() causes no additional write IO. It simply commits right now. Buffer cache simply delays the writes. The SD flash will see the same total number of writes regardless. Both things are undesirable with slower flash storage such as SD cards. Disabling fsync() and its cousins may improve performance with some workloads but it will not change flash cell life. I think disabling fsync() had more positive performance effect than tweaking vm.dirty_ratio. Both had strong positive effects for my system load case. Other cases will be uniquely different. YMMV. If you are using EXT3 with its insane journal everything mode, then yes, this would yield a big boost. If you're using XFS, it shouldn't make a huge difference, unless of course your workload is all metadata. Disabling fsync() in this manner is a very bad idea because it removes the safety features built into filesystems. If your UPS fails, the kernel crashes, etc, you will likely have a corrupt filesystem. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52e08b74.4000...@hardwarefreak.com
Re: Install Debian on SD
On 1/22/2014 9:12 PM, Артур Истомин wrote: On Tue, Jan 21, 2014 at 10:10:35PM -0600, Stan Hoeppner wrote: The OP will be using a browser. Browser cache writes, index updates, etc, will be far in excess of swap writes. If he uses Thunderbird (IceDove) with GLODA and offline caching enabled, that will produce even more writes. What is GLODA? http://lmgtfy.com/?q=gloda -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52e08c47.8070...@hardwarefreak.com
Re: Install Debian on SD
On 1/21/2014 5:19 PM, Bob Proulx wrote: James Kirk wrote: I would like to install debian on a SD (I've only a Company laptop and I would like to not modify the standard configuration). Sounds reasonable to me. Since SD is a flash memory and - as far as I know - it has a limited lifetime (in terms of I/O) not thought to run an OS, The quality of the SD depends upon the vendor. And unfortunately it is a market for lemons in that it is difficult for consumers to know which are better than others. But AFAIK most SD cards do implement at least some internal spares and wear leveling. how should I setup the installation (i.e. choosing a suitable filesystem, etc.) and/or modify the standard distribution in order to have a fully-functional Debian and make it compatible with a SD lifetime? You might look at Raspberry Pis, BeagleBone Blacks, Cubieboards, and many other systems routinely run from SD cards. Read the install docs and blogs for those types of systems and see what tweaks they suggest. They have been running on SD cards for quite a while now. Personally I don't do a lot of tweaking. I have been running an RPI system off of an SD since January 2013 24x7 plus several others for the last six months 24x7 and haven't had any problems with them yet. I normally set up a tmpfs in /etc/default/tmpfs. I don't do anything special with /var/log although I know some people put it on a tmpfs too but I like to see the logs if there is a reboot. I do use the noatime flag in /etc/fstab. I keep a backup. Which because it is an SD card is very easy to make a full device snapshot every so often. Talking about tuning the Raspbery Pi folks set vm.swappiness=1 which I disagree with. I suggest using the Linux upstream default of vm.swappiness=60 or even higher. Proponents of disabling program swap by setting it low say that they never want to swap. I have the opposite viewpoint. The point is to use the existing ram most effectively. If that means using ram for file system buffer cache instead of parts of programs that aren't ever accessed again then swapping out unused program space is better than hogging ram with it. The OP will be using a browser. Browser cache writes, index updates, etc, will be far in excess of swap writes. If he uses Thunderbird (IceDove) with GLODA and offline caching enabled, that will produce even more writes. Worrying about swap write flash wear is a non issue given the write load generated by these other desktop apps. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52df44bb.6030...@hardwarefreak.com
Re: V Taller Internacional de Tecnologías de Software Libre y Código Abierto
On 1/17/2014 5:16 PM, Vicios wrote: Hi all! Forwarding message to Debian spanish users list. Regards. That was UCE, i.e. spam. You're an idiot. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52d9c153.8010...@hardwarefreak.com
Re: Installation
On 1/13/2014 9:52 PM, Azeem Abdul Azeem wrote: I want to install Debian on my vertual machine (windows 200 R2 Hyper-V). So i need a bootable ISO. So kindly send me the link where i can download the bootable iso and install in the system http://www.debian.org/CD/torrent-cd/ -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52d4baae.1050...@hardwarefreak.com
Re: Question
On 1/12/2014 1:25 PM, Cameron Murgatroyd wrote: Hi I've recently become a frequent user of debian and I have a question if I were to want to make a .deb package for a game hack and to do it I needed to delete some files from the users file system before my files go in how would I do it? The first thing you need to do is post exactly what you're trying to accomplish, in detail. Lots and lots of detail. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52d2fd05.6040...@hardwarefreak.com
Re: PXE install, without internet?
On 1/11/2014 4:40 AM, Артур Истомин wrote: On Sat, Jan 11, 2014 at 01:21:30PM +0530, Anubhav Yadav wrote: Hey folks, I managed to convince my staff to switch to debian from windoze, and they agreed. So I managed to install a PXE server and successfully booted debian installer simultaneously on 20 machines using dhcp server. But they required an active internet connection to fetch packages over a mirror, and the bandwidth was very low, so at then end of the day, only one PC was able to set up. So is there a way to boot an entire 4gb dvd-iso from a server, so that I can install it on PC connected on a network?? You can create local mirror of repos. http://www.debian.org/mirror/ftpmirror Yes, with 20 machines to install via PXE boot, setting up a local mirror is the best method by far. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52d125e1.9040...@hardwarefreak.com
Re: PXE install, without internet?
On 1/11/2014 5:57 AM, Scott Ferguson wrote: On 11/01/14 22:07, Stan Hoeppner wrote: On 1/11/2014 4:40 AM, Артур Истомин wrote: On Sat, Jan 11, 2014 at 01:21:30PM +0530, Anubhav Yadav wrote: Hey folks, I managed to convince my staff to switch to debian from windoze, and they agreed. So I managed to install a PXE server and successfully booted debian installer simultaneously on 20 machines using dhcp server. But they required an active internet connection to fetch packages over a mirror, and the bandwidth was very low, so at then end of the day, only one PC was able to set up. So is there a way to boot an entire 4gb dvd-iso from a server, so that I can install it on PC connected on a network?? You can create local mirror of repos. http://www.debian.org/mirror/ftpmirror Yes, with 20 machines to install via PXE boot, setting up a local mirror is the best method by far. Mirror means even packages *not* required, apt-caching allows a local repository of only the packages that are actually required, with the additional benefits of handling multiple releases and repostories. I was using the term mirror much more loosely here, in the old installer terminology. I.e. select your mirror. I.e. simply copy the source files from the DVD he already has and export the proper directory tree structure via an httpd. Maybe I should have said setup an alternate installation source instead. Doing this would likely be faster than sharing the DVD, as was mentioned, and having 20 clients seek the drive head to death, taking days for the installs to complete. Get it on a hard drive and it'll be much faster. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52d2028a.30...@hardwarefreak.com
Re: Number of Procs
On 1/9/2014 6:52 AM, Jeff Bauer wrote: On 01/08/2014 10:41 AM, Hudson Flavio Meneses Lacerda wrote: hudson@musix:~$ free -m total used free sharedbuffers cached Mem: 435412 22 0 2 74 -/+ buffers/cache:334100 Swap: 1057143914 In 2014, that's a pretty meager amount of memory. Going into swap tells a tale, too. If possible, consider adding memory. Before replacing the SO-DIMM with a larger one... The host has 512MB of RAM and apparently 64MB is peeled off for the integrated GPU, with the remainder of the ~77MB difference apparently eaten by BIOS shadowing, etc. The first thing I would do is jump into the BIOS config and reduce the GPU memory to the minimum allowed, if it will go lower than 64. The OP doesn't currently have enough RAM to run compiz or openGL apps and is surely not currently doing so. If he is, that's the source of the problem right there. If this laptop allows a shared GPU memory size as low as, say, 4MB, that would free 60MB more RAM for Linux. That may not be enough to completely solve the slowness issues, but it should surely help to some degree. If the lowest allowed in 32MB that's still an additional 32MB for Linux. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52cf5be3.1020...@hardwarefreak.com
Re: why linux can't see my memory
On 1/5/2014 6:24 AM, Long Wind wrote: BIOS see all memory and Windows XP can run Yes, memory test OK after the memtest probe method To grub config add the mem parameter, e.g. kernel /boot/vmlinuz-3.2.X ro root=/dev/sdXX mem=1024M -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52c9c0ae.6010...@hardwarefreak.com
Re: Dovecot *requires* MySQL?
On 1/2/2014 9:04 PM, Jordan Metzmeier wrote: On Thu, Jan 2, 2014 at 8:34 PM, Bob Bernstein poo...@ruptured-duck.com wrote: Setting out to install dovecot-imapd on a squeeze host via apt-get, I discovered that: The following extra packages will be installed: dovecot-common libmysqlclient16 libpq5 mysql-common I understand that neither of the two mysql debs named above provide MySQL _server_ function, so why are they needed? I refuse to have a MySQL server on (or doing business with) any host for which I am responsible. I ma not really happy with that 'client' mysql thingie, to be completely honest! Dovecot does not require mysql. The dovecot-common package *recommends* dovecot-mysql. Apt installs recommended packages by default, but they are not required. You can exclude recommended packages with --no-install-recommends. The only technical downside to having mysql-common and libmysqlclient16 is about 5KB of disk space. Dovecot also supports LDAP and sqlite as the user database, so you should have seen dependencies for libldap and libsqlite as well, and also libpam which is the default user database. You'll also see libssl, openssl, etc. These are installed by default so that they simply work when you configure dovecot to use them, instead of pulling your hair out when the errors pile up in the log, and auth doesn't work. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52c63464.2010...@hardwarefreak.com
Re: moving buffers/caching from RAM to SSD
On 1/1/2014 12:43 AM, Muhammad Yousuf Khan wrote: Thanks all, it was probably be my misunderstanding with both technologies (RAM and SSDs). i have been told that SSDs are fast enough like RAMs and to show the performance that person refer his OS boot time which was dramatically minimized. i am also obsessed by the magic of ZFS. and ZFS people recommend using ZIL and logging on SSD. which apparently seems like the same process how linux cache things in RAM. Caching is used at many levels in computer systems, both in hardware, the OS kernel, and in software applications. But the implementations are quite different. Conceptually, Linux page caching and ZFS caching may be similar, but they are quite different in implementation, execution, and function. however with your help and searching on google i learn that there is a hell lot of difference in speed of RAM an SSD. Yes, a massive difference. The DRAM bandwidth of a basic desktop system today w/dual channel ddr3-1333 is ~20 GB/s. The fastest SSDs are ~0.5 GB/s, or 40x slower. The latency of modern DDR3 DRAM is ~50 ns (cache miss latency, not CAS cycle time). The latency of SSDs is ~100 μs, or 2000 times slower. Thanks all. i really appreciate your help. On Tue, Dec 31, 2013 at 10:33 PM, Stan Hoeppner s...@hardwarefreak.comwrote: On 12/31/2013 7:54 AM, Muhammad Yousuf Khan wrote: i dont know why i am saying is even practical or not. here is my free command @thor:# free -g total used free sharedbuffers cached Mem:31 31 0 0 0 26 -/+ buffers/cache: 3 27 Swap: 93 0 93 as you can see 27GB is being used in caching. i have few 160GB SSDs. can i move this buffers/caching load to my SSD. so that things could work more better. Muhammad, By design, the Linux kernel will use nearly all free memory for caching disk blocks and filesystem metadata when the memory isn't needed by other processes. When a process needs memory, the kernel simply drops some of the cached pages, freeing them for immediate use. This process takes a few tens of nanoseconds per 4KB page--it is instantaneous. It is because these pages can be freed instantly that Linux eats up all the RAM for cache. Cached file access is hundreds of times faster than disk access, even if disk is SSD. What you are seeing is the expected Linux kernel behavior. There is nothing wrong here, nothing to fix. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52c2ffeb.7010...@hardwarefreak.com -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52c42a1c.6060...@hardwarefreak.com
Re: moving buffers/caching from RAM to SSD
On 12/31/2013 7:54 AM, Muhammad Yousuf Khan wrote: i dont know why i am saying is even practical or not. here is my free command @thor:# free -g total used free sharedbuffers cached Mem:31 31 0 0 0 26 -/+ buffers/cache: 3 27 Swap: 93 0 93 as you can see 27GB is being used in caching. i have few 160GB SSDs. can i move this buffers/caching load to my SSD. so that things could work more better. Muhammad, By design, the Linux kernel will use nearly all free memory for caching disk blocks and filesystem metadata when the memory isn't needed by other processes. When a process needs memory, the kernel simply drops some of the cached pages, freeing them for immediate use. This process takes a few tens of nanoseconds per 4KB page--it is instantaneous. It is because these pages can be freed instantly that Linux eats up all the RAM for cache. Cached file access is hundreds of times faster than disk access, even if disk is SSD. What you are seeing is the expected Linux kernel behavior. There is nothing wrong here, nothing to fix. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52c2ffeb.7010...@hardwarefreak.com
Re: Distro upgrade
On 12/19/2013 9:54 AM, Verde Denim wrote: Just checking in to get a consensus - I'm running a (very stable) Debian workstation with Wheezy that has a lot of development/testing applications installed. I tried an install of hopper disassembler that threw errors concerning GLIBC and libffi so I contacted the package maintainers who told me that they setup an environment with Debian Jessie and the package works as expected. I'd like to work with this package, but am a curious to know if upgrading the release from Wheezy to Jessie might cause other issues with the large number of packages already running well. Any input, as always, is greatly appreciated. Why take any chances? Install/run Jessie and the application in a virtual machine. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52b319d1.4070...@hardwarefreak.com
Re: Wasted space in hard disk partitioning (was Soliciting Hardware Recommendations)
On 12/16/2013 10:12 PM, Stephen Powell wrote: On Mon, 16 Dec 2013 08:13:16 -0600, Stan Hoeppner wrote: Cool. Make sure you partition the SSD so that your first, and all, partitions start on a 4KB boundary. Many guides are available for your favorite partitioning tool. Linux does all IO in 4KB pages including filesystem and swap. If you don't align to 4KB you may get excessive erase cycles on the SSD, lowering performance, and cell life. Most folks seem to start the first SSD partition at sector 2048, which falls immediately after the first 1MB of the device. 1,048,576/4,096=256. So your first partition will start at 4KB page 257, if you will. If you create multiple partitions, make sure the size of each is evenly divisible by 4096 bytes, or they won't be aligned. That's good to know. But I really don't have much control over this when I partition using the Debian installer. ... Then boot a suitable live CD/DVD first and partition the SSD manually with parted. E.g. # parted (parted) unit s (parted) mkpart primary 2048 206847 (parted) set 1 boot on (parted) mkpart primary linux-swap 206848 2303999 (parted) set 2 swap on (parted) mkpart primary 2304000 212019199 (parted) quit This should give you a 100 MB partition for /boot, a 1 GB partition for swap, and a 100 GB partition for the root filesystem and everything else. All should be 1 MB aligned if I did my math and off-by-one-starting-from-zero correctly. With /boot and swap toggled on, the installer should use these automatically. You should only need to tell it which filesystem to format with, and to stick the root filesystem and everything else on the 100GB partition. Using the above you will have some unallocated free space at the end of the device. This is GOOD to have with an SSD. The controller will automatically use these unallocated cells for garbage collection and wear leveling. This will further extend the life of all cells of the drive. ... As you can see, something, somewhere, has done a number on me. The only boundary requirement is that the starting sector be a multiple of 8 to make 4096-byte boundaries. So why start the first partition at sector number 2048? ... The issue isn't simply with 4KB sector alignment, but also the erase block size of the SSD. Some use 128KB, some 512KB, some 1MB. But they don't tend to publish this information. By aligning to 1MB and making all partitions divisible by 1MB, you guarantee you're aligned to the erase block size. I won't go into any more detail on this as it's been thoroughly covered by others. Google erase block. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52b00a19.7060...@hardwarefreak.com
Re: Soliciting hardware recommendations
On 12/15/2013 5:44 PM, Stephen Powell wrote: On Sun, 15 Dec 2013 15:07:33 -0500 (EST), Stan Hoeppner wrote: I was figuring you'd probably do the install from a USB stick and I threw the DVD drive in as an afterthought. So yes, the ASUS burner above would be perfect. Over 4,000 sold and a 5 egg rating. Can't get any better than that. Good. I just ordered everything exactly as you recommended, except for the CD/DVD drive, for which I substituted the R/W model. To install Debian, I usually use the netboot installation image burned to a CD-R. Cool. Make sure you partition the SSD so that your first, and all, partitions start on a 4KB boundary. Many guides are available for your favorite partitioning tool. Linux does all IO in 4KB pages including filesystem and swap. If you don't align to 4KB you may get excessive erase cycles on the SSD, lowering performance, and cell life. Most folks seem to start the first SSD partition at sector 2048, which falls immediately after the first 1MB of the device. 1,048,576/4,096=256. So your first partition will start at 4KB page 257, if you will. If you create multiple partitions, make sure the size of each is evenly divisible by 4096 bytes, or they won't be aligned. If you decide to go this route and you run into any trouble getting it going, don't hesitate to email me off list. I may not respond quickly but I'll respond. That's a very generous offer. Thank you very much. I hope I won't need to take you up on that, but it's nice to know that I can if I need to. It would be very poor form of me to recommend a bunch of parts and then run away when problems arise during integration. :) This is a bit different than recommending a model of TV or microwave... You know, I almost added an anti-static wrist strap to the order, but in the end decided not to. My basement is so humid, even in winter, that I doubted it would be a problem. However, my dehumidifier, which normally runs 9 or 10 months out of the year, is currently showing the relative humidity at 30% (and it's off, set for 50%), so maybe I'll get one anyway. I'm a bit of a hypocrite here. I never use straps myself but always recommend it to others. I've been doing computer work for a couple of decades+, and I always do assy/repair on metal benches. Before I handle a component I grab the bench, and while I'm working I keep a foot or knee against a bench leg or the table top frame. It's simply 2nd nature. And it negates the restriction of un/clipping a strap if I need to leave the bench. Thanks again, Santa. Ho, ho, ho, you're welcome. :) This will be my first truly new system since 1994. Wow, you should be in for a treat. Especially with that SSD. You're going to need to watch your console closely. Even big operations will complete before you hear the click from the Enter key. ;) I've been using other people's throw-aways and used systems since then. But nobody I know is throwing away 64-bit systems. Not yet anyway. Maybe not throwing them away, but there are tons of decent used 64 bit x86 boxen on Ebay. But even at those prices one can build a better system for little more money, such as yours. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52af0a7c.1020...@hardwarefreak.com
Re: Soliciting hardware recommendations
Hi Stephen, Have you been eagerly anticipating my reply to your hardware thread? I hope I don't disappoint. :) On 12/14/2013 7:23 AM, Stephen Powell wrote: On Sat, 14 Dec 2013 02:39:04 -0500 (EST), David Christensen wrote: ... I'd like it to have a usable CSM, so I can continue to run my favorite boot loader, LILO. CSM = IBM Cluster Systems Management? No, CSM = Compatibility Support Module, a feature of UEFI-compliant motherboards that provides a PC-compatible BIOS for booting legacy BIOS operating systems, such as the LILO boot loader. This is incompatible with Connected Standby Mode, whose initials are also, unfortunately, CSM. Connected Standby Mode is apparently a requirement for Windows 8 certification. Connected Standby Mode must be disabled in order for a Compatibility Support Module to be enabled. There is no need for CSM if you purchase a fairly modern board yet designed prior to the UEFI revolution. Here is a complete list of quality DIY parts that will meet your stated needs, including a non-UEFI motherboard: Gigabyte GA-78LMT-S2P(rev 5.0) AM3+ AMD 760G/SB710, AMD 3000 GPU, DVI AMD FX-4300 3.8GHz quad core CPU, 2x 2MB L2 cache, 4MB L3 cache 2x 4GB DDR3-1333 Crucial DDR-SDRAM modules Samsung 840 EVO 120GB SSD - #1 SSD in quality, performance, and $/GB ASUS DVD - it's DVD to boot a net install image, if needed, decent unit Apevia mATX cube case, 500W PSU, excellent airflow, I own one, good case Hewlett Packard 23 Widescreen HD monitor, 1920x1080, DVI $ 52 http://www.newegg.com/Product/Product.aspx?Item=N82E16813128504 $110 http://www.newegg.com/Product/Product.aspx?Item=N82E16819113287 $ 60 http://www.newegg.com/Product/Product.aspx?Item=N82E16820148347 $ 90 http://www.newegg.com/Product/Product.aspx?Item=N82E16820147247 $ 19 http://www.newegg.com/Product/Product.aspx?Item=N82E16827135304 $ 87 http://www.newegg.com/Product/Product.aspx?Item=N82E16811144140 $120 http://www.newegg.com/Product/Product.aspx?Item=N82E16824176250 $538 Total Upon totaling this I'm really surprised such a quality, high performance machine with a good 23 LCD can be had for less than $550, especially with a Samsung 120GB SSD. Some of the prices above are holiday specials. I didn't check the normal prices. So if you like this combo you may want to check the expiry dates on the specials and order beforehand. I verified the FX-4300 should work in this board so you should have no surprises. You may need to flash it so the BIOS identifies the CPU correctly, but the board should boot and work fine even without doing so. See: http://www.gigabyte.com/support-downloads/cpu-support-popup.aspx?v=AE72F6F9D4ADD0D7995989AF720CB4D0FF2976CBFB7DB4DE70BB3E832BEDD48Ds=Socket%20AM3+cs=AMD%20760G FYI, I searched for more than an hour trying to find an AMD dual core CPU plus mobo combo that would be suitable. I ran into two problems here. First, all of the current dual core AMD CPUs lack L3 cache and have somewhat tiny 1MB L2 caches. As with virtual machines, emulation workloads tend to perform better with larger caches due to context switches, TLB shootdowns, etc. So this model FX-4300 with 1MB L2 per core and 4MB shared L3 should perform better for you, even if you can't make use of all 4 cores immediately. Everyone else buying 4/6/8 core desktop CPUs is in the same boat, whether they know it or not, so don't sweat it. The industry has decided to take the multi-core path which users simply cannot yet follow, because most developers aren't yet threading their applications. Second, all of the socket FM2 motherboards which take the dual core chips are built/certified for Windows 8, and have UEFI BIOS. Again this mobo is non-UEFI, so I think this combo will be a better solution for you all around Stephen. Motherboard manual to read before you buy (I always do): http://download.gigabyte.us/FileList/Manual/mb_manual_ga-78lmt-s2p_v.5.0_e.pdf I hope this is the type of complete, concise information, recommendation, you were looking for Stephen. -- Stan TheHardwareFreak -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52adbd13.9060...@hardwarefreak.com
Re: Soliciting hardware recommendations
On 12/15/2013 10:56 AM, Stephen Powell wrote: On Sun, 15 Dec 2013 09:30:43 -0500 (EST), Stan Hoeppner wrote: ... Wow, you even managed to get a monitor thrown in and still make the ~$500 target. I was expecting to pay around $500 for the system only, with the monitor costing extra. Yeah, like I said, it surprised me a bit when I tallied it, considering I made a point to select only top quality gear. The thing that really surprised me was getting a fast quad core CPU, the Samsung 120GB SSD, and the 23 HP LCD in there, and still sneaking in under $550. This will be a pretty awesome system for the price. I'm jealous. :) ... I see just one problem. I don't see any indication of R/W capability in the CD/DVD drive. It appears to be read-only. I want to be able to burn install images downloaded from the internet. However, I did find this item http://www.newegg.com/Product/Product.aspx?Item=N82E16827135204 for just a buck more. Do you see any problem substituting this for the one you suggested? I was figuring you'd probably do the install from a USB stick and I threw the DVD drive in as an afterthought. So yes, the ASUS burner above would be perfect. Over 4,000 sold and a 5 egg rating. Can't get any better than that. ... Makes sense. And thanks for all the time that you have put into this. You're welcome. If you decide to go this route and you run into any trouble getting it going, don't hesitate to email me off list. I may not respond quickly but I'll respond. You said it's been a while since you built a system, so add a static strap to your order and use it, if you don't already have one. Winter is static charge season, and the last thing you want is to have your board not work after accidentally sending static discharge into it. Often folks don't even realize it happened and curse their DOA board as being junk, etc. ... I hope this is the type of complete, concise information, recommendation, you were looking for Stephen. It is. You did not disappoint. On that note, prepare for potential disappointment with the cutesie little temperature LCD on the Apevia case. Mine was DOA. I wasn't about to RMA the entire case to Newegg and wait a week or more for a replacement just to get the temp LCD fixed. The system was up and running and immediately in production. I bought the case for the superior airflow and the form factor, not the LCD display. Out of curiosity, I contacted Apevia who offered to ship me a new LCD panel at no charge, but they said I had to pay the shipping, ~$8, or RMA it through Newegg. Not seeing the CPU heatsink temp (nowhere close to core temp anyway) and the HD case temp simply wasn't an issue for me. In your system, the SSD will run ever so slightly above T_case, so you won't even bother to hook the HD temp probe to the SSD. If yours works I'd hook up the CPU heatsink probe and I'd connect the HD probe to the Northbridge heatsink. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52ae0c05@hardwarefreak.com
Re: Maximum RAM
On 12/12/2013 2:42 AM, Scott Ferguson wrote: Perhaps because 64-bit gives their use case brings disadvantage but no advantages? Perhaps for other reasons. To assume that you *should* use 64-bit in all cases is incorrect. There are old 32 bit PAE only machines around with plenty of capability today, albeit with a lot more space and power consumption for the performance. Take the Unisys ES7000 Orion 230 for example. It's a 32 processor Foster Xeon 72 mainframe, 32 bit PAE only CPUs. Original price when new in 2001 was ~$300,000 USD. Today? A few thousand, if you could find a complete working unit at a surplus equipment dealer or on Ebay. CPU count: 32 CPU type: Intel Xeon MP Foster, first gen NetBurst Specs: 1.4-1.6 GHz, 256KB L2, 1MB L3 System cache: 256MB static RAM, 32MB per 4 CPU module, 8 modules System RAM: 64GB ECC SDRAM, 128x 512 MB DIMMs, 32-way interleaved RAM bandwidth: 20 GB/s sustained, 25.6 GB/s peak IO Slots: 64 PCI 2.1 66 MHz 32 PCI 2.1 33 MHz IO Bandwidth: 5 GB/s sustained There are a number of workloads at which this 13 year old PAE only system would offer excellent performance today. It would make one heckuva server for web, mail, database, etc. It would have decent SETI or Folding throughput though individual work unit processing would be pretty slow compared to today's CPUs. You'd be hard pressed to find a 32 bit PCI FC400/800 or SAS controller, but Intel still sells a PCI GbE card. This allows for MPIO iSCSI over multiple HBAs and GbE links to modern iSCSI SAN RAID arrays. Say 16 HBAs, 4 links to each of 4 arrays with 24x 2.5 SAS drives, 96 drives total, 3.2 GB/s throughput. This is obviously an obscure and unlikely scenario, but it is a good example of why one would choose to run a PAE kernel. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52a9a5c3.2090...@hardwarefreak.com
Re: Hardware Question about RAM and Capacitors
On 12/12/2013 1:11 PM, Goren Buckwalk wrote: ... I have a system with two AMD Athlon 2400 MP processors and the motherboard has 4 slots for RAM. That makes this board ~10 years old. ... About a month ago, I found the box crashed again and beeping on reboot, so going through the same elimination process found only the stick in the first slot would work. It seemed odd two sticks could go bad at the same time, so I tried them in that first slot and both worked. I tried adding some back and no matter the combination except for one single stick, always got a beep-fest. So I think all the RAM is good, but the last three slots are bad (or maybe just the 2nd is bad, and then the 3rd and 4th can't work without the 2nd??). DIMM sockets don't fail. What fails is the power circuit supplying voltage/current to the sockets. Looking closer I see three capacitors on the board (nearer to the CPUS than the ram slots) have a rusty looking coating on the top. One is worse looking than the other two (eh?), the two have some part of their tops that still look shiny silver. I've heard of capacitors going bad on motherboards, but never saw any. Is this rust a sign of failure and could their failure be the cause of the RAM slot problems? Possibly. If you look at the PCB below the caps you'll see a colored residue there as well (gravity). The caps get old, get hot, and the aluminum casing separates. When this happens the oil spills out and the cap is done/dead. This assumes a horizontal case. If the case is vertical then gravity will pull the oil along the bottom edge of the cap cylinder wall and pool there before drying. It may drip before drying as well, so look at components nearby underneath the split caps. Most mobo manufacturers switched about 5-6 years ago to using all solid caps, no more oil/film parts. These have a much longer lifespan. Other than the lockups/memory beeping I've never really seen any issues with this system (did replace bad drive once or twice). It has boinc/seti/asteroids, webserver(s) and other stuff running (except when running BOINC its not heavily stressed but generally is doing _some_ work pretty much all the time). Are three rusty capacitors and 3 bad slots just a coincidence? Thanks. Probably not coincidence. But keep in mind other components may have failed as well. You could try recapping the board using your soldering skills and one of these kits, or purchase individual caps for your board: http://www.badcaps.net/pages.php?vid=21 But given the cost of $25-30 and your time, the fact that recapping may not be a complete fix, and the gear is 10 years old, you may be better off buying new guts. This seems to currently be the least expensive Newegg AMD based combo: $ 37http://www.newegg.com/Product/Product.aspx?Item=N82E16813138376 $ 43http://www.newegg.com/Product/Product.aspx?Item=N82E16819103888 $ 26http://www.newegg.com/Product/Product.aspx?Item=N82E16820231150 $106Mobo, CPU, two DIMMs Even with a single core at 2800 MHz clock the Sempron will run circles around your twin MPs at 2x 2000 MHz clock, due to the 1 MB L2 cache (vs 256KB), the integrated memory controller, and dual channel DDR3-1333. The Athlon MP 2400+ Socket A bus throughput is 2.1 GB/s or roughly 1.05 GB/s per CPU. This Sempron combo? 21 GB/s, 10x faster. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52aaba58.3000...@hardwarefreak.com
Re: Postfix - run script from email
On 12/7/2013 6:58 AM, Tony van der Hoff wrote: I also have a number of remote, normally unattended, locations, with dynamic IPs, containing IP cameras to keep an eye on things. I'm not able to access/change the camera firmware. These cameras send me an email if their IP changes, containing the new IP address. I think you're attacking this from the wrong angle. Whether these cameras are on a private RFC 1918 network or a public broadband network, you should be able to accomplish your goals without need of SMTP emails. If broadband you must have a DSL/cable modem/router in front of the camera with public IP on the provider side and private IP on the inside. Use dynamic DNS on the public interface to get name-IP resolution of the modem/router. On the router, configure the DHCP server to only assign one IP address, or a range that starts and ends with the same address. Create a NAT rule for the camera IP/port. From then on access the camera by fqdn, not IP. If this is a totally private network you can assign sticky static addresses in your DHCP server so that each IP camera always receives the same address on lease renewal. No email required. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52a3ad96.8080...@hardwarefreak.com
Re: My apologies
On 12/4/2013 5:18 AM, Chris Bannister wrote: On Tue, Dec 03, 2013 at 04:52:06PM -0600, Stan Hoeppner wrote: Many threads will wonder a bit and everyone accepts this. Just try to Agreed! Also ... wander. :) Those damn homonyms... Heheh, just caught it myself on a re-read of my posts after seeing Guntner para-quote me as saying expects when I said accepts. This is bound to happen especially with words one uses so infrequently-- I can't recall the last time I used wander in text or speech. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/529f9801.8070...@hardwarefreak.com
plese stop Re: A rookie's query: Want to about Debian and the related
This thread began on Nov 24th, 10 days ago. There have been 211 posts (including this one) in this thread. I dare say it ceased being productive or insightful many, many posts ago. And it ceased having anything to do with Debian Linux quite a while ago. Let's move on to something worthwhile. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/529e416d.2000...@hardwarefreak.com
Re: plese stop Re: A rookie's query: Want to about Debian and the related
On 12/3/2013 2:45 PM, Ralf Mardorf wrote: On Tue, 2013-12-03 at 14:39 -0600, Stan Hoeppner wrote: This thread began on Nov 24th, 10 days ago. There have been 211 posts (including this one) in this thread. I dare say it ceased being productive or insightful many, many posts ago. And it ceased having anything to do with Debian Linux quite a while ago. Let's move on to something worthwhile. Mailer used by Stan is Mozilla/5.0 (Windows NT 5.1; rv:24.0) Gecko/20100101 Thunderbird/24.1.1, so filtering unwanted threads doesn't work that good for Thunderbird, perhaps a drawback of Thunderbird when using mailing lists. Just a guess ;). I'll start depositing paper bags full of dogs poop on your front porch 20 times a day and light each on fire. Your reaction to that will clearly demonstrate whether the problem has anything to do with your ability to filter the flaming dog poop, or if something needs to be done about the person depositing it on your porch. Especially given that rules are already in place to prevent such a thing. In this case they simply haven't been enforced yet. For your sake, take a clue, and extricate yourself from this argument before they are enforced. Don't assume that simply because it doesn't happen often that people don't get banned from debian-user. Ralf, do you really want to be known as the only person in 2013 to be banned from debian-user? And for what? A stupid argument over mail clients? -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/529e4b7f.2020...@hardwarefreak.com
Re: My apologies
On 12/3/2013 4:06 PM, Ralf Mardorf wrote: no doubt about it, many of my mails were unneeded, but the most mails for the thread, some from me, many from others, were absolutely ok. Agreed, and thank you Ralf. If I should write again too much off-topic, please send me a note off-list. If a newbie does ask something that is not Debian related enough for your taste, please consider that this is a user list. Maybe some expert talks belong to a developers list ;). Many threads will wonder a bit and everyone accepts this. Just try to realize when you've gone off the reservation. If you don't, and someone asks to end the thread, without naming names or assigning blame, you've got to take the opportunity being provided to you and bow out of the thread gracefully. This way everyone wins, everyone saves face. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/529e6096.9020...@hardwarefreak.com
Re: raid problem
On 11/29/2013 4:43 PM, François Patte wrote: Bonsoir, I have a problem with 2 raid arrays: I have 2 disks (sdc and sdd) in raid1 arrays. One disk (sdc) failed and I replaced it by a new one. Copying the partition table from sdd disk using sfdisk: sfdisk -d /dev/sdd | sfdisk /dev/sdc then I added the 2 partitions (sdc1 and sdc3) to the arrays md0 and md1: mdadm --add /dev/md0 /dev/sdc1 mdadm --add /dev/md1 /dev/sdc3 There were no problem with the md0 array: cat /proc/mdstat gives: md0 : active raid1 sdc1[1] sdd1[0] 1052160 blocks [2/2] [UU] But for the md1 array, I get: md1 : active raid1 sdc3[2](S) sdd3[0] 483138688 blocks [2/1] [U_] And mdadm --detail /dev/md1 returns: /dev/md1: Version : 0.90 Creation Time : Sat Mar 7 11:48:30 2009 Raid Level : raid1 Array Size : 483138688 (460.76 GiB 494.73 GB) Used Dev Size : 483138688 (460.76 GiB 494.73 GB) Raid Devices : 2 Total Devices : 2 Preferred Minor : 1 Persistence : Superblock is persistent Update Time : Fri Nov 29 21:23:25 2013 State : clean, degraded Active Devices : 1 Working Devices : 2 Failed Devices : 0 Spare Devices : 1 UUID : 2e8294de:9b0d8d96:680a5413:2aac5c13 Events : 0.72076 Number Major Minor RaidDevice State 0 8 510 active sync /dev/sdd3 2 002 removed 2 8 35- spare /dev/sdc3 While mdadm --examine /dev/sdc3 returns: /dev/sdc3: Magic : a92b4efc Version : 0.90.00 UUID : 2e8294de:9b0d8d96:680a5413:2aac5c13 Creation Time : Sat Mar 7 11:48:30 2009 Raid Level : raid1 Used Dev Size : 483138688 (460.76 GiB 494.73 GB) Array Size : 483138688 (460.76 GiB 494.73 GB) Raid Devices : 2 Total Devices : 2 Preferred Minor : 1 Update Time : Fri Nov 29 23:03:41 2013 State : clean Active Devices : 1 Working Devices : 2 Failed Devices : 1 Spare Devices : 1 Checksum : be8bd27f - correct Events : 72078 Number Major Minor RaidDevice State this 2 8 352 spare /dev/sdc3 0 0 8 510 active sync /dev/sdd3 1 1 001 faulty removed 2 2 8 352 spare /dev/sdc3 What is the problem? And how can I recover a correct md1 array? IIRC Linux md rebuilds multiple degraded arrays sequentially, not in parallel. This is due to system performance impact and other reasons. When the rebuild of md0 is finished, the rebuild of md1/sdc3 should start automatically. If this did not occur please let us know and we'll go from there. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/529979f8.1050...@hardwarefreak.com
Re: MIT discovered issue with gcc
On 11/22/2013 7:34 PM, Andrew McGlashan wrote: http://www.securitycurrent.com/en/research/ac_research/mot-researchers-uncover-security-flaws-in-c the team ran Stack against the Debian Linux archive, of which 8575 out of 17432 packages contained C/C++ code. For a whopping 3471 packages, STACK detected at least one instance of unstable code. So 3471 Wheezy packages had one ore more instances of gcc introduced anomalies. And the kernel binary they tested had 32. As an end user I'm not worried about this at all. But I'd think developers may want to start taking a closer look at how gcc does its optimizations and creates these anomalies. If the flaws are serious they should obviously takes steps to mitigate or eliminate this. I didn't read the full paper yet, but I'm wondering how/if the optimization flag plays a part in this. I.e. does O2 produce these bugs but OO (default) or Og (debugging) does not? -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52911cb9.5010...@hardwarefreak.com
Re: Installation of Wheezy on a Fujitsu Primergy Raid1
On 11/18/2013 5:03 AM, Joerg wrote: Hi Joerg, I've read all the responses to your post. The reason they contain questions and no answers is because you've provided insufficient detail about your system and its RAID hardware. I have a Fujitsu primergy TX100 S1 and installed a bios-raid1 (LSI raid from Fujitsu - This is the first example of insufficient information. You need to provide the model# of the Fujitsu/LSI HBA. 2 disks prepartitioned into 1 primary and 5 extension disks each - done by ctrl-M). This is a technically incorrect description of what you have, and is the reason for everyone's confusion. You need to be technically accurate here. It sounds like what you have done is a create a RAID1 group of two physical disks in the LSI firmware, and then on that RAID1 device created 6 Virtual Disks. Each virtual disk is exposed as a SCSI LUN. Thus, the Linux kernel, the Wheezy installer, are seeing 6 disks. Is this correct? If not, please explain in detail what virtual devices you are exposing from the LSI HBA. This machine used to have Lenny but an important component was destroyed so I decided to proceed to Wheezy skipping squeeze (which I do not have as netinst disk!!!) When I tried to install wheezy via netinst from CD I could not get out of the partition menue for installation of basic components ... The parted-menue shows: 1. Raid (the whole disk - no partitions so far) 2. the 1st scsi-device with all partition on that device (contains important data which must not be destroyed) 3. the 2nd scsi-device with all partition on that device (contains important data which must not be destroyed) 1st and 2nd scsci devices are partitioned the same way from the previous Lenny installation. My question: How can I proceed with the LSI raid 1 to get a functioning system again? Provide the LSI physical/virtual disk configuration information requested above and we'll go from there. However, the act of going through the steps required to gather said information should inform you as to which device:partition the Lenny installation resides on. Once you know that device:partition number, the rest should be straightforward. Simply tell the Wheezy installer to put its files on that partition. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/528b0676.5020...@hardwarefreak.com
Re: Wich port should I use for Xeon E3-1230v2 processor ?
On 11/15/2013 10:48 AM, John Ostrowski wrote: Hi, I recently bought Dell PowerEdge T110 II with Intel Xeon E3-1230v2 Processor (3.3GHz, 4C/8T, 8M Cache). I don't know which port I should use. At the moment I am using Debian 7.1.0 AMD64 and I AMD64 is the correct architectural port for all Xeon Processors going back at least 8 years. have serious issue. Desktop keeps locking all the time. I can move mouse but system does not react. Best regards. John Ostrowski The problem is most likely your graphics driver. Which GPU is in this Dell? -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52869a27.7050...@hardwarefreak.com
Re: Questions about new hardware Debian (or Linux in general)
On 11/13/2013 9:22 AM, Dan Ritter wrote: ftp://download.nvidia.com/XFree86/gpu-hdmi-audio-document/gpu-hdmi-audio.html Dan, you're awesome. I bet alot of nVidia users, especially MythTV users, will find this immensely helpful. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52843eb5.6090...@hardwarefreak.com
Re: Questions about new hardware Debian (or Linux in general)
On 11/12/2013 5:37 PM, Jon N wrote: ... There is one an area that I'm pretty unsure of. I am planning on purchasing a Nvidia video card and disabling the built in Intel video support. Since I plan to use this computer as a MythTV frontend/backend (as well as for general web browsing/email) getting the audio out on the Nvidia card's HDMI port is important to my particular setup. So will the audio automatically be switched to the Nvidia cards HDMI connector? No, it won't be automatic. And frankly I don't believe nVidia supports HDMI digital audio pass through, nor any discrete GPU card. For argument's sake, let's say it does. Then you run into the problem that the onboard audio chip can't pass digital audio through PCIe to the nVidia HDMI port. None of them are designed to do this, that I'm aware of. If I were you I'd get a mainboard with with HDMI out and use the CPU's GPU. Mobos that have onboard HDMI have their audio chips wired to the HDMI port, the chips support PCM/AC3 digital output, and selecting the HDMI output for digital audio is pretty straightforward. The Intel GPU should be plenty powerful enough for HD1080 output. If you decide it's not, and want to add a discrete card, you'll need a mobo with coaxial digital SPDIF output, or Toslink optical digital output, and a TV or A/V receiver that is cable of using an HDMI input for video while using coax or Toslink for audio. Nearly all modern A/V receivers support this. WRT LCD/Plasma TVs I have no idea how many support this. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/5282c88a.3040...@hardwarefreak.com
Re: Questions about new hardware Debian (or Linux in general)
On 11/12/2013 7:11 PM, Doug wrote: On 11/12/2013 07:32 PM, Stan Hoeppner wrote: On 11/12/2013 5:37 PM, Jon N wrote: ... There is one an area that I'm pretty unsure of. I am planning on purchasing a Nvidia video card and disabling the built in Intel video support. Since I plan to use this computer as a MythTV frontend/backend (as well as for general web browsing/email) getting the audio out on the Nvidia card's HDMI port is important to my particular setup. So will the audio automatically be switched to the Nvidia cards HDMI connector? No, it won't be automatic. And frankly I don't believe nVidia supports HDMI digital audio pass through, nor any discrete GPU card. For argument's sake, let's say it does. Then you run into the problem that the onboard audio chip can't pass digital audio through PCIe to the nVidia HDMI port. None of them are designed to do this, that I'm aware of. If I were you I'd get a mainboard with with HDMI out and use the CPU's GPU. Mobos that have onboard HDMI have their audio chips wired to the HDMI port, the chips support PCM/AC3 digital output, and selecting the HDMI output for digital audio is pretty straightforward. The Intel GPU should be plenty powerful enough for HD1080 output. If you decide it's not, and want to add a discrete card, you'll need a mobo with coaxial digital SPDIF output, or Toslink optical digital output, and a TV or A/V receiver that is cable of using an HDMI input for video while using coax or Toslink for audio. Nearly all modern A/V receivers support this. WRT LCD/Plasma TVs I have no idea how many support this. I have an NVidia GeForce GTX 550 Ti card in a machine with on-board Realtek 662 sound decoder. Normally--that is locally--sound is decoded by the Realtek on the mobo. The NVidia card has a mini HDMI connector (I needed an adapter to standard) and the card has a sound decoder in it. This isn't the case. All digital audio (AC3, PCM, DTS, etc) is decoded at the endpoint device. In this case that is the TV or A/V receiver. Everywhere else in the path the digital audio stream is simply passed through. The trick with Linux is getting all of the devices recognized, and being able to select which 'path' the digital stream should take. Using Windows XP, I could run a movie thru HDMI to my TV set and picture and sound would come thru perfectly. NVidia provides a driver for Windows that makes this just work. And this is the key. nVidia registers an audio output device that can be selected in control panel as the preferred output device. WSS then directs digital audio through this device. There is no such equivalent in Linux, that I'm aware of. I never could get the sound to work to the TV from my PCLinuxOS system, but obviously, it's THERE! (The video comes thru the TV OK.) One contributor to the PCLOS forum who seemed to know what he was doing tried very hard to make the sound work, but without success. (Ideally, it should be possible to get both sound and video _at the same time_ both locally and at the TV.) Just so you know. Which is why I recommended the mobo-down HDMI solution, which doesn't have these problems of output device selection. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/5282fcc7.9090...@hardwarefreak.com
Re: Questions about new hardware Debian (or Linux in general)
On 11/12/2013 10:09 PM, Jon N wrote: On Nov 12, 2013 7:32 PM, Stan Hoeppner s...@hardwarefreak.com wrote: On 11/12/2013 5:37 PM, Jon N wrote: ... There is one an area that I'm pretty unsure of. I am planning on purchasing a Nvidia video card and disabling the built in Intel video support. Since I plan to use this computer as a MythTV frontend/backend (as well as for general web browsing/email) getting the audio out on the Nvidia card's HDMI port is important to my particular setup. So will the audio automatically be switched to the Nvidia cards HDMI connector? No, it won't be automatic. And frankly I don't believe nVidia supports HDMI digital audio pass through, nor any discrete GPU card. For argument's sake, let's say it does. Then you run into the problem that the onboard audio chip can't pass digital audio through PCIe to the nVidia HDMI port. None of them are designed to do this, that I'm aware of. Wow, I'm glad I asked that question :-). If I understand this correctly it it doesn't matter if there is any video hardware on the mainboard, in the processor, or none at all, you still can't get sound from a audio chipset on the mainboard to the video card's HDMI connector anyway. That's not actually how it works. The OS directs digital audio. In the case of a mobo audio chip, digital audio is sent to this chip no matter what, and it either decodes it to analog and dumps it out the discrete analog ports, or it passes the digital stream unmolested through a digital connector, either HDMI, coax, or Toslink. In this case, there are discrete mobo traces from the audio chip to each of these 3 outputs. You select the audio output in the driver. The chip then passes it to the connector you choose. In the case of sending digital audio out the HDMI connector on a graphics card, this is done purely in software, and the data stream never goes to the audio chip. It's sent from the application through the audio driver directly to the video card HDMI drive chip. The reason I mentioned going from the audio chip to the HDMI of the vid card is that, AFAIK, the Alsa driver doesn't support multiple digital out devices, and neither nVidia/ATI support HDMI audio output in their drivers. If you use a mobo audio chip, its driver should allow you to select any of the audio ports on the board. You may have to do it statically in a config file instead of selecting it on the fly as in MS Windows, but you should be able to use any connector on the back panel nonetheless. So again, if you want to send video+audio over a single HDMI cable to the TV, I think this is your only option. If I were you I'd get a mainboard with with HDMI out and use the CPU's GPU. Mobos that have onboard HDMI have their audio chips wired to the HDMI port, the chips support PCM/AC3 digital output, and selecting the HDMI output for digital audio is pretty straightforward. I think pretty much all the Mobos have HDMI in them, especially since they support a processor line that all has built in video. Not by a long shot. Only about half the boards available at Newegg have HDMI. And, obviously, all boards lacking integrated GPUs, or supporting AMD/Intel performance CPUs with no GPU, do not have HDMI. These are your high end SLI/X-fire boards. These have integrated audio chips but their digital out is limited to Toslink/coax. This is no problem in MS Windows as you can select the HDMI output on the discrete GPU board. With Linux, thus far, it appears one is SOL. I was planning on Nvidia simply because a) I use it now and b) I am under the impression that they have better overall support (i.e.: just work better). But I may be underestimating how well the built in Intel video solution works. And it would same me money by not purchasing a new board, use less electricity (love that), and maybe even make the system quieter (no fan on a separate video card). I will check over at the MythTV mailing list about it. 2D video processing takes almost no GPU horsepower at all. Any modern CPU w/integrated GPU can handle broadcast HDTV or Blue Ray HD video without breaking a sweat, and so can any modern mobo GPU. It's 3D gaming where they have weakness, especially at high resolutions and high texture detail. But if you're only doing video, any of them is more than sufficient. The Intel GPU should be plenty powerful enough for HD1080 output. If you decide it's not, and want to add a discrete card, you'll need a mobo with coaxial digital SPDIF output, or Toslink optical digital output, and a TV or A/V receiver that is cable of using an HDMI input for video while using coax or Toslink for audio. Nearly all modern A/V receivers support this. WRT LCD/Plasma TVs I have no idea how many support this. I currently use my DVI out to HDMI in on my receiver, and s/pdif for the audio, it works fine. I thought it would be nice to have it all in one. I It is nicer. And it's a couple of clicks to make
Re: harddrive spins down and up again after 5s
CC'ing back to the list. On 11/11/2013 7:42 AM, patrick wrote: Am 11.11.2013 00:01, schrieb Stan Hoeppner: On 11/10/2013 3:18 PM, Ralf Mardorf wrote: On Sun, 2013-11-10 at 14:19 -0600, Stan Hoeppner wrote: Thus the problem may not be caused by the operating system. It is the OS, it's gvfs. Sure, it doesn't cause the spin down, but gvfs does cause the unwanted spin up. You may be correct. However, do note that Patrick's only indication that his drive was spinning down/up every 5 minutes was that he heard it. Which is why I was starting from ground zero with the troubleshooting. Modern drives, both laptop and desktop, are so quiet as to be inaudible to many/most people. Thus, what Patrick is hearing may not even be the hard drive, but could very well be his CPU or GPU fan, or his DVD drive motor spinning up/down. oh yes, i didnt specify the hardware. its a toshiba sattelite pro c660-2tq, its about one year old but got barely usage on it. im sure that the cpu fan sounds different and is located in a different position. the cpu fan is incredibly loudactually the laptop should not have a pro in its name. theres no dvd/cd in the drive im working with pcs for over 10 years now, i always build them from the ground up so im sure that the harddrive is what im listening to. Ok, so you're pretty sure it is the drive cycling down/up. hdparm -I /dev/sda says its a: toshiba MK5075GSX barely google hits for this product. shouldnt gvfs show up in mount? also, it looks like gvfs is pretty essential to gnome 3, how much work did you do to create a replacement packet? Ralf will have to help you with the Gnome/GVFS aspect of this as I'm not familiar with it. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/5281644b.9050...@hardwarefreak.com
Re: harddrive spins down and up again after 5s
On 11/10/2013 1:46 PM, patrick wrote: my freshly installed debian wheezy is spinning the harddrive up and down all the time... First let's establish this is actually the case. You didn't state new computer new drive or old computer/drive, new wheezy. If new computer/drive, which drive is this? WD Green series drive firmware auto parks the heads after ~5 seconds and spins the drive down automatically after 30s, IIRC. Thus the problem may not be caused by the operating system. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/527fea35.8020...@hardwarefreak.com
Re: harddrive spins down and up again after 5s
On 11/10/2013 3:18 PM, Ralf Mardorf wrote: On Sun, 2013-11-10 at 14:19 -0600, Stan Hoeppner wrote: Thus the problem may not be caused by the operating system. It is the OS, it's gvfs. Sure, it doesn't cause the spin down, but gvfs does cause the unwanted spin up. You may be correct. However, do note that Patrick's only indication that his drive was spinning down/up every 5 minutes was that he heard it. Which is why I was starting from ground zero with the troubleshooting. Modern drives, both laptop and desktop, are so quiet as to be inaudible to many/most people. Thus, what Patrick is hearing may not even be the hard drive, but could very well be his CPU or GPU fan, or his DVD drive motor spinning up/down. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52801052.2020...@hardwarefreak.com
Re: No space left on device (28) but device is NOT full!
On 11/4/2013 10:28 PM, Tazman Deville wrote: ... Got it! find . -name 'popularity-*' | xargs rm -rf (passes the files to rm one at a time). Glad you got it squared away Anthony. Normally I'd suggest filing a bug report against the problem application, but since the system is Squeeze it's probably already been addressed. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/5278a96e.7040...@hardwarefreak.com
Re: No space left on device (28) but device is NOT full!
On 11/5/2013 1:21 AM, Richard Hector wrote: On 05/11/13 16:51, Stan Hoeppner wrote: Second, you have a serious problem here because it is your root filesystem that has run out of inodes. You need to ask yourself why you have 1.7M files in your rootfs. That's very dumb. Or perhaps That's not generally advisable. or similar. Richard Or perhaps you should remove your politically correct inspired, sensitivity trainer ground, colored lenses so you can correctly read and comprehend what I typed. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/5278b1e4.4070...@hardwarefreak.com
Re: No space left on device (28) but device is NOT full!
On 11/4/2013 8:30 PM, Tazman Deville wrote: On Mon, Nov 04, 2013 at 06:43:45PM -0500, Hecber Cordova wrote: Hi, Did you check inodes usage? (df -i) I could be inodes availability rather than block availability. AHA! I have no idea what the significance of this is, but df -i gives $ df -i FilesystemInodes IUsed IFree IUse% Mounted on /dev/sda71729920 1729920 0 100% / So, yeah...inodes, but I'm ignorant of what that means, or how to resolve that. First it means you're using a filesystem with a small fixed number of inodes, obviously EXT. Second, you have a serious problem here because it is your root filesystem that has run out of inodes. You need to ask yourself why you have 1.7M files in your rootfs. That's very dumb. That's what /home and /data and other places are to be used for. To remedy this you will need to copy files off of the rootfs to another filesystem, then delete them from your rootfs to free some of these inodes. Food for thought: your /dev/sda7 is an EXT filesystem of 26GB with 1.7M inodes. XFS would give you ~23M inodes on a 26GB filesystem. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52786b3e.6020...@hardwarefreak.com
Re: Replacing failed drive in software RAID
On 11/1/2013 9:19 AM, Pascal Hambourg wrote: Stan Hoeppner a écrit : This is precisely why I use hardware RAID HBAs for boot disks (and most often for data disks as well). The HBA's BIOS makes booting transparent after drive failure. In addition you only have one array (hardware) instead of 3 (mdraid). MD RAID arrays can be partitionned, or contain multiple LVM logical volumes. So you don't have to create multiple arrays, unless they are of different types (e.g. RAID 1 and RAID 10 as in this thread). Yes, I'm well aware of md's capabilities. I was speaking directly to the OP's situation. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/5273d7b2.2000...@hardwarefreak.com
Re: Replacing failed drive in software RAID
On 11/1/2013 12:23 PM, Pascal Hambourg wrote: Stan Hoeppner a écrit : On 11/1/2013 9:19 AM, Pascal Hambourg wrote: Stan Hoeppner a écrit : This is precisely why I use hardware RAID HBAs for boot disks (and most often for data disks as well). The HBA's BIOS makes booting transparent after drive failure. In addition you only have one array (hardware) instead of 3 (mdraid). MD RAID arrays can be partitionned, or contain multiple LVM logical volumes. So you don't have to create multiple arrays, unless they are of different types (e.g. RAID 1 and RAID 10 as in this thread). Yes, I'm well aware of md's capabilities. I was speaking directly to the OP's situation. So was I. In the OP's situation, there are arrays of different types (1 and 10) so you cannot have one array even with hardware RAID. Of course he can, and it is preferable to use a single array. The only reason the OP has a separate RAID1 is the fact that it is much simpler to implement boot disk failover with md/RAID1 than with md/RAID10. And in fact that is why pretty much everyone who uses only md/RAID has at least two md arrays on their disks: a RAID1 set for dual MBR, /boot, and rootfs, and a separate RAID5/6/10/etc for data. With hardware based RAID there are no such limitations. You can create one array and throw everything on it. No manually writing an MBR to multiple disks, none of md's PITA requirements. Zero downside, lots 'o upside. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52745675.4030...@hardwarefreak.com
Re: Replacing failed drive in software RAID
On 10/31/2013 3:41 PM, Bob Proulx wrote: Veljko wrote: I'm using four 3TB drives, so I had to use GPT. Although I'm pretty sure I know what I need to do, I want to make sure so I don't loose data. Three drives are dying so I'm gonna replace them one by one. Sounds like a good plan to me. It is what I would do. It is what I have done before when upgrading sizes to larger sizes. This is what I plan to do: Replacing sda ... Did I overlook something? Will this going to work? Very well thought out plan! Looks okay to me. I like it. Some boot issues to discuss however. Is this a BIOS boot ordering boot system booting from sda? In which case replacing sda won't have an MBR to boot from. You can probably use your BIOS boot to select a different disk to boot from. And then after having booted install grub on the other disk. (Sometimes the BIOS boot order will be quite different from the Linux kernel drive ordering.) I am unfamiliar with the sgdisk backup and load-backup operation. I am not sure that will restore the grub boot sector. This isn't too scary because you can always boot one of the other drives. Or boot a debian-install rescue media. But after setting up the replacement disk it will probably be necessary to install grub upon it in order for it to be bootable as the first BIOS boot media. And very often I have found that a second disk that I thought should have had grub installed upon it did not and when removing sda I find that the system won't grub boot from sdb. Therefore I normally restore sda, boot, install grub on sdb, then try again. But if you know ahead of time you can re-install grub on sdb and avoid the possible hiccup there. But if you are concerned about writes to sdb then I would simply plan to boot from the debian-installer image in rescue mode, assemble the raid, sync, then replace sdb, and repeat. You can always install grub to the boot sectors after replacing the suspect disks. Hopefully this makes sense. This is precisely why I use hardware RAID HBAs for boot disks (and most often for data disks as well). The HBA's BIOS makes booting transparent after drive failure. In addition you only have one array (hardware) instead of 3 (mdraid). You have only 3 partitions to create instead of 9, these residing on top of the one array device, not used to build multiple software array devices. So you have one /boot, root fs, and data, and only one MBR to maintain. The RAID controller literally turns your 4 drives into one, unlike soft RAID. The 4 port Adaptec is cheap, $200 USD, and a perfect fit for 4 drives: http://www.adaptec.com/en-us/products/series/6e/ http://www.newegg.com/Product/Product.aspx?Item=N82E16816103229 And because it has 128MB cache you get a small performance boost. I was also thinking about inserting one drive and copying data from RIAD to it so I have backup if something goes wrong. Would that be right thing to do, or that would just load drives unnecessarily and accelerate their failure? Are you asking about the one drive inserted being large enough to do a full system backup? If so then I think it is hard to argue against a full backup. I think I would do the full backup even with the extra disk activity. It is read, not write, and so not as bad as normal read-write disk activity. Agreed. In which case you might consider that instead of replacing all disks one by one that you could simply do a full backup, then create the new system with lvm and raid as desired, and then restore the backup onto the newly constructed partitions. After you have the full backup then your original drives would be shut off and available as a backup image too in that case. So that also seems a very safe operation. This is my preferred method. Cleaner, simpler. Still not as simple as moving to hardware RAID though. Or since you have four new drives go ahead and construct a new base configuration with the four new drives with lvm+raid as desired. And then clone directly from the old system disks to the new system disks. Then boot the new system disks. This has much more offline time than the replace one disk at a time that you outlined above. I normally do the sync one disk at a time since the system is online and running services normally during the sync. But there are many ways to accomplish the task. And yes there is more down time with this method. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/5272d469.8010...@hardwarefreak.com
Re: How and where lspci utility gathers information about hardware components?
On 10/27/2013 1:18 PM, Martin T wrote: lspci utility shows information regarding devices on various buses like PCI or PCI Express. For example on IBM ThinkPad T42 laptop: [snip] Am I correct that lspci uses SMBus which is present both on PCI and PCI Express? In addition, how are the model numbers detected? SMBUS is not required. Once the PCI driver loads it scans all devices on the PCI bus or host bridge controller, collecting the PCI device vendor/model data during this enumeration process, along with programming/initializing each device, and then populates /sys/bus/pci/devices/:00:xx.x/vendor /sys/bus/pci/devices/:00:xx.x/device which are read by lspci. For example Intel Corporation PRO/Wireless 2915ABG Wi-Fi adapter. Is there a small non-volatile memory on each device which stores information about particular hardware device? No, these IDs are hard coded in PCI device registers, which is why each ID is exactly 4 bytes in length. PCI ID is part of the PCI standard, part of the PCI Configuration Space. lspci cross references the vendor and device IDs in /sys/bus/pci/../ with the PCI ID table to give you the detailed device information. See: http://pciids.sourceforge.net/ Example: http://pci-ids.ucw.cz/read/PC/8086 -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/526d641a.1060...@hardwarefreak.com
Re: Hp Proliant 460c drivers
On 10/21/2013 2:57 PM, Erick Ocrospoma wrote: Is this your first rodeo with HP servers? I have installed Debian 7.1 on a few hp proliant blade servers, after installing I've seen that network is not working properly. It recognises around 8 interfaces, mii-tool says that there is not any link present. Post your dmesg and lspci output WRT to the NIC hardware so we can verify which Ethernet chip it is. I've searched on Hp support page and they only bring a small iso for hdd drivers. Does anybody has had the same trouble with this model? They are new servers, firmware is not yet updated (is scheduled for tomorrow). Which BL460C? The G8? The BL460C been on the market since 2010 at the latest, 3 years ago. There have been at least 3 hardware generations of this blade. However... The NC373i has existed since at the latest 2007, 6 years ago. It uses a Broadcom chipset. It is included in just about every HP server, has been for over half a decade. It is fully supported by Linux going back to 2.6. Googling, i've read that there's only official support for RHEL family. Anyway, we prefer using Debian rather than other distro. Hope it is possible. This is probably simply a non-free firmware issue. The Wheezy installer probably didn't install the firmware by default, nor ask you to do so. See: https://wiki.debian.org/Firmware Look for: Broadcom NetXtreme II BCM5706/5708/5709/5716 Driver Broadcom NetXtreme II BCM57710/57711/57711E/57712/57712_MF/57800/57800_MF/57810/57810_MF/57840/57840_MF Driver -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52667f40.4090...@hardwarefreak.com
Re: Hp Proliant 460c drivers
On 10/22/2013 10:11 AM, Erick Ocrospoma wrote: On Oct 22, 2013 8:36 AM, Stan Hoeppner s...@hardwarefreak.com wrote: On 10/21/2013 2:57 PM, Erick Ocrospoma wrote: Is this your first rodeo with HP servers? Blades yes I have installed Debian 7.1 on a few hp proliant blade servers, after installing I've seen that network is not working properly. It recognises around 8 interfaces, mii-tool says that there is not any link present. I'm pretty damn sure that blade doesn't have 8x 10GbE interfaces. It should have one, or two if you purchased the dual port mezzazine board. Post your dmesg and lspci output WRT to the NIC hardware so we can verify which Ethernet chip it is. lspci | grep Eth Ethernet Controller: Emulex corporation One connect 10GB NIC (be3) (rev01) The kernel driver is be2net. Is the driver loading? dmesg output please. Are you using static assignment or DHCP? What is the content of /etc/network/interfaces? -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/5266b295.4030...@hardwarefreak.com
Re: Hp Proliant 460c drivers
On 10/22/2013 1:42 PM, Erick Ocrospoma wrote: On Oct 22, 2013 12:15 PM, Stan Hoeppner s...@hardwarefreak.com wrote: On 10/22/2013 10:11 AM, Erick Ocrospoma wrote: On Oct 22, 2013 8:36 AM, Stan Hoeppner s...@hardwarefreak.com wrote: On 10/21/2013 2:57 PM, Erick Ocrospoma wrote: Is this your first rodeo with HP servers? Blades yes I have installed Debian 7.1 on a few hp proliant blade servers, after installing I've seen that network is not working properly. It recognises around 8 interfaces, mii-tool says that there is not any link present. I'm pretty damn sure that blade doesn't have 8x 10GbE interfaces. It should have one, or two if you purchased the dual port mezzazine board. Post your dmesg and lspci output WRT to the NIC hardware so we can verify which Ethernet chip it is. lspci | grep Eth Ethernet Controller: Emulex corporation One connect 10GB NIC (be3) (rev01) The kernel driver is be2net. Is the driver loading? dmesg output please. Are you using static assignment or DHCP? What is the content of /etc/network/interfaces? I have find the correct interfaces, from the group of 8 the two first ones are the NICs. There are only asigned only two virtual NICs per bay. I have just used the install cd1 so this must be a generic driver, anyway. Now, the problem is with the HBA fibber adapter, there's a SAN on the datacenter which Debian should recognise. Emulex brings official drivers for CentOS and due to this servers will be under high load, should I follow the official ones? We are testing under CentOS too, works fine but in both cases, Debian and CentOS, now we need to make work the fibber adapter. Hardware drivers are part of the Linux kernel. They are not distribution specific. You didn't post dmesg output so I don't know if the Emulex FC HBA driver is loading or not. I'd guess it is. The first thing you need to do is enter the HBA BIOS during POST and see what LUNs are being unmasked to your HBA WWNs, if any. Read the HBA manual to learn how to do this. If the switches have not been programmed adding the WWNs of the FC HBAs to an appropriate zone, and if the SAN array administrator has not exposed LUNs to the HBA WWNs, you won't see any LUNs in the HBA BIOS, nor any /dev/sdX devices that map to SAN LUNs. You literally won't see anything on the SAN. I can tell this is your first rodeo with FC SAN technology. There is nothing automatic with FC SAN hardware. Zones, paths, security, ports, etc, must typically all be manually configured on each piece of the SAN, including the array controller, switches, and HBAs. Most vendor introductory SAN courses are 3-5 days, and most offer 3 course levels, beginner, intermediate, and advanced. If you've had no prior SAN training or self study, you may be in for a rough ride, unless you have a SAN admin, or someone on staff with knowledge. If so you will need to work closely with that person to get this working. FWIW, I taught myself enough in about 15 hours to integrate my first SAN using Qlogic switches and HBAs, IBM and Nexsan disk arrays, using strictly online vendor documentation, and Google. So it is possible to fly by the seat of one's pants with one's first SAN deployment, but I am not your average tech, and most people can't pull this type of thing off working alone. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52674101.8010...@hardwarefreak.com
Re: postfix
On 10/21/2013 6:55 AM, Jeremy T. Bouse wrote: On 21.10.2013 07:56, Pol Hallen wrote: Hey all :-) I'm sorry for banal question but I didn't find any answer to my question. In the /etc/postfix/main.cf I see many parameters like: smtpd_sasl_auth_enable = yes but also: smtp_sasl_auth_enable = yes So, what is the difference of smtpd_parameter and smtp_parameter? (smtpd and smtp) Thanks for help Pol The smtpd_* prefixed commands are Server configurations where Postfix is operating as the server providing the service. The smtp_* prefixed commands are Client configurations where Postfix is operation as a client talking to another server. In many fewer words: smtpd = inbound smtp = outbound -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52653c2e.1040...@hardwarefreak.com
Re: xtrs z80
Beco rcb at beco.cc writes: Is xtrs broken under wheezy? What is it anyway? A Tandy Trash 80 emulator? -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/loom.20131015t093704-...@post.gmane.org
Re: xtrs z80
On 10/15/2013 1:47 PM, Beco wrote: On 15 October 2013 04:38, Stan Hoeppner s...@hardwarefreak.com wrote: Beco rcb at beco.cc writes: Is xtrs broken under wheezy? What is it anyway? A Tandy Trash 80 emulator? Hi Stan, Yep, its an emulator for TRS-80 ! Please, don't call it by its ugly nickname (Trash!) :) I say that with affection. I cut my computing teeth on the TRS-80 Model III and IV. I taught myself BASIC, Pascal, and Assembly Language on these machines in the 2nd half of the 80s, in my early teens. I'm having lots of fun with this emulator. Today I could install a harddrive, mind you! It has (I believe, I need to compute it yet) 13 MB!! I'll put all software from various disketts I've collected in this single file. Now I just need to make it boot from the HD and I don't need even the OS on drive :0 anymore. As soon as I put a c compiler on it, I'll start using only my TRS-80 to work. Bye debian! It was nice to be with you! But things evolves, you know! I don't know if C was even available for the TRS-80 back then. If it was it didn't cross my RADAR. Ahaha! I wonder what my students would do when I ask for the next assignment: -Do a program that reads 2 numbers and calculates the average, in C. - Just it, teacher? - Yep. But on a TRS80. LOL. Aww, don't make it so easy. Have them do it in assembler. :) -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/525e0556.2020...@hardwarefreak.com
Re: Unable to locate mail spool file.
On 10/12/2013 9:27 AM, Paul Cartwright wrote: ok, I have exim4 setup as internet site, using Maildir. my /home/user/Maildir is there. I sent a test message and a new message shows up in /home/user/Maildir/new . in Thunderbird when I try to check mail in that account is says Unable to locate Mail spool file. It is a UNIX Movemail account. What am I doing wrong? I just want to get my local linux account email in IMAP in Thunderbird.. /var/log/exim4/mainlog shows Completed, so it is doing it's job, so I guess I am missing something in the Thunderbird setup? Thunderbird 24.0 No, what you're missing in an IMAP server. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52599b4f.6060...@hardwarefreak.com
Re: Ethernet bonding mode 5 only using one Slave adapter.
On 10/11/2013 2:42 AM, Muhammad Yousuf Khan wrote: [Cut]. Are dual and quad port Intel NICs available in your country? Not very easily but yes, we can arrange. i personally have PCIe 4 Port intel NIC. so this can be arranged. I recommend Intel NICs because they simply work, every time, full bandwidth, full Linux kernel support, great feature set, etc. Very high quality, long lasting. I had an Intel Pro/100 in service in an MX mail server for over 10 years. Still works. ... just a very basic question i am into virtualization for few years on Debian box. i never host a VM on external box. i have more then 10 nodes and all VMs are hosted on local Mdadm RAID drives. Just to have an idea. if you like to suggest. how many VM can be hosted on 1G link. i know your next statement will be it depends upon the utilization of your VM and decision would be made on IO stats basis Yes, it does depend on exactly that. but just asking in general how many general VMs can be hosted on 1G LAN that are more or less untouched throughout the day. If idle? As many as you can fit in memory up to the hypervisor limit, or virtual IP address limit, if there is a limit on these. It's possible to create VMs that have no network stack at all. In that case there is literally no limit WRT the shared GbE link. and my big big time confusion is backup the VM from Virtualization terminal. lets say for a while 2 VM are running on 1GB link and i am taking a backup of a VM from virtual server. as the server is connected to external storage on 1 GB link. first virtual server will bring all the virtual drive data from External box to local RAM via same 1GB link on which VMs are hosted. it does mean that when backup will start all other VMs has to suffer? so even if 1 VM is running and we are making/creating a backup then how can we avoid chocking the link or bottle neck. Ok, so apparently I misunderstood previously. I was under the impression that you had an NFS storage server box, a backup server box, and many physically boxes on which you were running virtual machines. I.e. 6 or more computers connected to a GbE switch. If I understand correctly now, all of your VMs are on one PC, and there is an NFS server somewhere on the network where you store the files. Is there a switch between the PC with all of the VMs, and the NFS server? If so... There are a couple of ways to address this: 1. Add another GbE interface on the PC and dedicate it to NFS traffic. You should be able to bind the NFS client to a specific IP address. This will require setting up source based routing so NFS traffic only uses the new interface. Without source based routing Linux will always use the first bound adapter for all outbound traffic. This dedicates the current NIC to everything other than NFS traffic, so the VMs have 1Gb/s for non-NFS traffic, and 1Gb/s for NFS traffic, 2Gb/s aggregate. This would be my preferred method. It's low cost, just a NIC and a cable. But you have a steep learning curve ahead WRT Linux routing. A bonus is you'll learn a lot about Linux networking in the process. 2. Implement QOS features in the switch, if it has them, to limit the amount of bandwidth used by NFS traffic. The problem with this method is that most switches don't allow this on a per port basis, but on a VLAN basis. Which means you'd be limiting NFS bandwidth everywhere, network wide, not just to the VM PC. ... any howto document on DRBD and GFS2 on debian? as i am using debian and only debian in overall environment. DRBD+GFS2 has got a native support on Redhat (as GFS2 is owned by Redhat). i do not have the experience nor confidence on stability of the both. i will be glad if you share any specific one with Debian. DRBD and GFS2 are both kernel modules. Their configuration on Debian should be little different than on any Linux distro. i found this http://pve.proxmox.com/wiki/DRBD http://www.drbd.org/users-guide/ch-gfs.html the above is Primary/Primary installation means both drbd drives can be mounted. but there is a question. if i can mount in Primary/Primary mode on both the nodes then what is the need of GFS? just asking for my learning. The key word here is mount. Linux cannot mount a block device. DRBD is a block device. Linux mounts filesystems. Filesystems reside on top of block devices. No two hosts can mount the same filesystem on a shared block device unless it is a cluster filesystem. Cluster filesystem are designed specifically for this purpose. However, in the real world, the block device under GFS2 and OCFS2 filesystems is most often a LUN on a fiber channel or iSCSI SAN storage array, not DRBD. ... Thanks for sharing such a detail and very helpful email. You're welcome. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact
Re: Ethernet bonding mode 5 only using one Slave adapter.
On 10/9/2013 5:51 AM, Muhammad Yousuf Khan wrote: [cut]... What workload do you have that requires 400 MB/s of parallel stream TCP throughput at the server? NFS, FTP, iSCSI? If this is a business requirement and you actually need this much bandwidth to/from one server, you will achieve far better results putting a 10GbE card in the server and a 10GbE uplink module in your switch. Yes, this costs more money, but the benefit is that all client hosts get full GbE bandwidth to/from the server, all the time, in both directions. You'll never achieve that with the Linux bonding driver. I appreciate your detailed email. it clears lots of confusion going inside my mind. the reason of increasing bandwidth is testing clustering /VM hosting on NFS and VM backups. My company is about to host their product inside our foreigner office-premises and i will be maintaining those servers remotely,. therefore i need to consider high availability of our service and that's why trying to test different technologies that can full-fill our requirement. Specificity testing Ceph clustering for hosting purposes. and for backing up my VMs as you know VMs are huge and moving them around on 1GB x-over point to point link takes time. so i thought i could increase some of the bandwidth and can use link aggregation to avoid single point of failure. i agree with you on buying 10GB LANs but unfortunately as i am testing this stuff very far far away from US these cards are not easily available in my country, thus unnecessarily expensive. Are dual and quad port Intel NICs available in your country? if you still have any advice for such scenario as mine i will be more glad to have it. Before a person makes a first attempt at using the Linux bonding driver, s/he typically thinks that it will magically turn 2/4 links of Ethernet into one link that is 2/4x as fast. This is simply not the case, and is physically impossible. The 802.3xx specifications do not enable nor allow this. And TCP is not designed for this. All of the bonding modes are designed first for fault tolerance, and 2nd for increasing aggregate throughput, but here only from one host with bonded interfaces to many hosts with single interfaces. There is only one Linux bonding driver mode that can reliably yield greater than 1 link of send/receive throughput between two hosts, and that is balance-rr. To get it working without a lot of headaches requires a specific switch topology. And its throughput will not scale with the number of links. The reason for this is that you're breaking a single TCP session stream into 2 or 4 streams of Ethernet frames each carrying part of a single TCP stream. This can break many of the TCP stack optimizations such as window scaling, etc. You may also get out of order packets, depending on the NICs used, and how much buffering they do before generating an interrupt. Reordering of packets at the receiver decreases throughput. Thus each link will have less throughput than when running in standalone mode. Most of the above information is covered in the kernel document. The primary driving force you mentioned behind needing more bandwidth is backing up VM images. If that is the case, increase the bandwidth only where it is needed. Put a 4 port Intel NIC in the NFS server and a 4 port Intel NIC in the backup server. Use 4 crossover cables. Configure balance-rr and tweak bonding and TCP stack settings as necessary. Use a different IP subnet for this bonded link and modify the routing table as required. If you use the same subnet as regular traffic you must configure source based routing on these two hosts and this is a big PITA. Once you get this all setup correctly, this should yield somewhere between 1-3.5 Gb/s of throughput for a single TCP stream and/or multiple TCP streams between the NFS and backup servers. No virtual machine hosts should require more than 1 Gb/s throughput to the NFS server, so this is the most cost effective way to increase backup throughput and decrease backup time. WRT Ceph, AIUI, this object based storage engine does provide a POSIX filesystem interface. How complete the POSIX implementation is I do not know. I get the impression it's not entirely complete. That said, Ceph is supposed to dynamically distribute data across the storage nodes. This is extremely vague. If it actually spreads the blocks of a file across many nodes, or stores a complete copy of each file on every node, then theoretically it should provide more than 1 link of throughput to a client possessing properly bonded interfaces, as the file read is sent over many distinct TCP streams from multiple host interfaces. So if you store your VM images on a Ceph filesystem you will need a bonded interface on the backup server using mode balance-alb. With balance-alb properly configured and working on the backup server, you will need at minimum 4 Ceph storage nodes in order to approach 400
Re: Ethernet bonding mode 5 only using one Slave adapter.
On 10/8/2013 4:41 AM, Muhammad Yousuf Khan wrote: i am using bond mode balance-alb. and here is my /etc/network/interfaces ... auto bond0 iface bond0 inet static address 10.5.X.200 netmask 255.255.255.0 newtork 10.5.x.0 gateway 10.5.x.9 slaves eth2 eth3 #bond-mode active-backup bond-mode balance-alb bond-miimon 100 bond-downdelay 200 ... note : as you can see in /proc/net/bonding/bond0 file the active link is eth2 and the bwm-ng showing transmission is also on eth2. even i use two session i thought it could work like round robin as in mode 0 but both sessions are transmitting data from eth2. With balance-alb, packet distribution across links is dictated by the receiving system's bond interface logic, not the sending system. The receiving system uses ARP replies to trick the sending host's interface into transmitting to one of the receiver's multiple physical interfaces. Thus, when balance-alb receive isn't working, it's usually due to this ARP reply trickery not working correctly. See balance-alb in https://www.kernel.org/doc/Documentation/networking/bonding.txt what i want to achive is per packet load balancing. This is not possible with mode balance-alb. The best you can get with balance-alb is per session load balancing, not per packet. If you want per packet transmit load balancing you must use balance-rr, but balance-rr does not balance receive traffic. if i send two packets it moves out from both link eth2 and eth3. so i can combine 4x1 LAN card and achieve 4 GB of transmit rate and redundancy. The only way to achieve this is between two hosts both with the same port count and both using balance-rr, either with x-over cables, or a dumb (non-managed) switch. i know 4 GB network output achievement depends upon the hardware quality. i Physical link throughput by the hardware is the least of your worries. The major hurdle is getting the bonding driver working the way you want it to. If not, the majority of the time you'll only get 1 GbE link of throughput. will check on that too. but at least bwm-ng could show me packet activity on all the links not only active link. Read the kernel doc I provided. My guess is that in the test case for which you provided numbers, only one slave on the receiving system was active. Follow the examples in the kernel doc and you should be able to straighten this out and achieve performance closer to your goal. What workload do you have that requires 400 MB/s of parallel stream TCP throughput at the server? NFS, FTP, iSCSI? If this is a business requirement and you actually need this much bandwidth to/from one server, you will achieve far better results putting a 10GbE card in the server and a 10GbE uplink module in your switch. Yes, this costs more money, but the benefit is that all client hosts get full GbE bandwidth to/from the server, all the time, in both directions. You'll never achieve that with the Linux bonding driver. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52551987.1080...@hardwarefreak.com
Re: Building computer
On 10/4/2013 4:44 AM, Paul Cartwright wrote: On 10/03/2013 05:11 PM, Stan Hoeppner wrote: No apology necessary. I'm on many open lists (LKML) where hitting reply-to-list only goes to the sender. So I've been guilty myself a few times. so that is list specific... I wondered, because sometimes I hit reply it goes to the person, other times it goes to the list.. Thunderbird.. This isn't an issue with TBird, or any MUA. The clients simply obey/honor the list headers. For example: X-Mailing-List: debian-user@lists.debian.org archive/latest/657649 List-Id: debian-user.lists.debian.org List-Post: mailto:debian-user@lists.debian.org Precedence: list The List-Post: header contains the list posting address. When you hit reply-to-list in TBird this is the address it selects for populating the To: field in the reply. If you reply to a message sent from a listserver that does not provide a List-Post: header, then the address in the Reply-To: header is inserted into the To: field of the reply. Ergo, reply-to-list only works if a List-Post header is present. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/524f42b7.3050...@hardwarefreak.com
Re: Building computer
On 10/3/2013 7:49 AM, Lisi Reisz wrote: On Thursday 03 October 2013 05:52:06 Stan Hoeppner wrote: In my early 40s A mere strippling! And there was I pegging you as a sage*. ;-) Lisi * a teacher venerable for years, and of sound judgment (Wiktionary) I'm sorry, I'm not [The One]. I'm just another guy. -- Neo to Trinity, The Matrix P.S. Sorry Stan - I did not mean to send to you personally. I am trying to train myself not to click on reply for the list, but not being very successful. Old dogs and new tricks? ;-( No apology necessary. I'm on many open lists (LKML) where hitting reply-to-list only goes to the sender. So I've been guilty myself a few times. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/524ddd74.4000...@hardwarefreak.com
Re: Arduino TRE
On 10/3/2013 2:55 PM, Beco wrote: Hi guys, Anyone have tested this Arduino TRE [1] with Debian? I'm very happy such enterprises still believe in, and produce hardware to linux! :) I bet this 1-GHz Sitara AM335x processor is very fast, suitable for most end-users to do basic stuff, like a small office or something. Cheers, Beco. [1] http://blog.arduino.cc/2013/10/03/a-sneak-preview-of-arduino-tre/ You need to ask on the Debian ARM mailing list, or search the archive. Subscribe: debian-arm-requ...@lists.debian.org Archive: http://lists.debian.org/debian-arm/ -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/524ddfe6.9080...@hardwarefreak.com
Re: Building a computer for compiling and CPU emulation (Re: Building computer)
On 10/1/2013 9:16 AM, Joel Rees wrote: On Tue, Oct 1, 2013 at 7:00 AM, Stan Hoeppner s...@hardwarefreak.com wrote: On 9/29/2013 6:01 AM, Joel Rees wrote: ... http://en.wikipedia.org/wiki/List_of_AMD_mobile_microprocessors tells me that AMD A4-1250, AMD A6-1450, and E2-1800 all have AMD-V. The E2-1800 has a half-Meg L2 cache, but higher speed CPU (1.7GHz). The A4-1250 has 1 Meg L2 cache but lower speed CPU (1.0 GHz). The A6-1450 has 2 Meg L2 cache to share between twice the CPUs. The A4-1450 is probably the best choice, not taking cost into consideration. Here's why: 1. Turbo core up to 1.4GHz 2. All 3 models have 512KB L2/core, no advantage for any 3. Temash core has better IPC than Zacate 4. 4 cores @ 1.4GHz should give better compile times than 2 cores @ 1.8GHz. 4 @ 1GHz should as well. -j4 or higher 5. A6-1450 is 9W chip, E2-1800 is 18W chip, longer run time 28nm vs 40nm The one downside is that for non-compute intensive operations, such as normal interactive GUI apps, say PDF viewing, browser rendering, etc, it may be considerably slower than the 1.8GHz E2-1800, due to the 800MHz clock deficit, as the turbo core may not kick in a lot here. And WRT turbo core, I'm not quite sure what this means: Selected parts support Turbo Dock technology, that can increase CPU and GPU frequencies when external cooling is available. So if the unit you purchase doesn't have a variable speed fan that can fulfill this requirement, this may mean you can't get 1.4GHz turbo mode. And I'm just guessing that devices of this class may not include forced air cooling. Sorry I don't have all the answers here, but maybe this helps get you a bit closer. One thing I can assure you of is that for your stated use case, IIUI correctly, all of these CPUs are very likely woefully inadequate for the task. Yes and no. Most of the compiling will be much smaller than a kernel, not even complete packages. I think I said it but you've clipped that part, but this is a course in programming that includes writing some drivers. Got it. Yeah, compiling a few hundred or thousand lines of code shouldn't be that bad. Emulating the superH processor is going to be a bit demanding, Instruction set level emulation is going to be extremely demanding if you intend to run a complete emulated OS environment. If you strictly use it for things like executing and debugging individual subroutines, code segments, etc, it shouldn't be too bad. particularly if I find myself wanting to compile a superH kernel and not having access to the school labs over a long holiday. I am aware of that and will plan accordingly. Yep. Mostly, I was trying to dig up the AMD-V support and something in the thread pointed me the right direction. Having AMD-V or Intel VT isn't an absolute requirement. Both will speed up context switches. As important as the new thread switching hardware in these CPUs is the size of the TLBs. CPUs without virtualization support tend to have inadequate TLBs. When you switch between VMs you end up flushing most of the TLB entries and reloading them. CPUs with large L2/L3 caches help mitigate this to a degree. Which is one of the reasons you see ginormous caches on server oriented CPUs, Xeon and Opteron, up to 34MB combined L2/L3. Most server workloads are transaction oriented and throughput is dictated by network/disk latency. A huge cache does nothing for you here. But when you run many virtual machines it provides serious benefit. This is one of the reasons I recommended going for the largest cache you can get. Unfortunately there's not much difference in cache sizes in the processor class you're looking at. Have you considered a refurb or used laptop? The Core2 Duo SL9600 offers 2.13GHz clock, 6MB L2 cache, at 17W. The Mobile PhenomII P650 2.6GHz clock w/2MB L2, 25W. I'm not planning getting more than AMD-V in a portable machine. Don't want to carry a boat battery with me. :-p Heheh, love the deep cycle reference. Though if you're like many mobile users, in the States anyway, most of the time a wall plug isn't far away, in either distance or time. Never visited Japan so I can comment on the situation there. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/524bc151.9070...@hardwarefreak.com
Re: Building computer
On 10/1/2013 5:13 PM, Catherine Gramze wrote: On Sep 30, 2013, at 10:33 PM, Stan Hoeppner s...@hardwarefreak.com wrote: Actually they were, up to the point you finally told us what screen resolution you use. That changes things quite a bit, or I should say changed one thing dramatically. I recommended a fast dual core CPU because more cores will be wasted. The i3-4340 3.6GHz Haswell would have been as fast as the quad core you bought, and saved ~$40-50. 4GB RAM, more than 4GB is wasted, but as I said previously it's cheap so buy more if you want, won't hurt. An SSD, and APU graphics. The only change in that recommendation, now that I know your screen resolution, is shifting to a very fast high bandwidth discrete card. 2560x1440 is a pretty insanely high gaming res if you want high frame rates and smooth rendering at high detail. Most hard core gamers wouldn't touch 2560x1440 without SLI/Xfire. If going w/a single reasonably priced card you're going to want/need a model with a 384 bit bus. Extremely high resolutions require extremely high memory bandwidth. The 384bit nVidia models are all above $600. The AMD 7950s can be had in the low $200s, and the 7970s in the low $300s. Both are 384bit. You are right about the quad core not making any real difference. I have run the activity monitor and observed how one core is at 70% and another at 22% and two others unused. Average user applications are not yet written to take advantage of multiple core processors. Nor will they ever be. I'm glad you looked into it and see it now. If you search the list archives you'll see I've written a bit on the multi-core issue, made the dual core recommendation many times, and went so far as to suggest a new dual core design that AMD or Intel could produce that would benefit everyone. But they're shackled to multi-core now because their marketing of the past 6 years or so has tried to convince everyone that more cores are better. Why? Because they're unable to increase per core performance at a rate which justifies buying new CPUs. If they did they'd have to reverse their marketing message, and that is just fraught with problems. The problem of efficient parallel programming will be with us for a VERY long time. And it's not that programmers, specifically game programmers, aren't crafty or smart enough to write parallel code. They are far more capable than others. The problems is that only a finite amount of a given program's flow logic can be parallelized. The rest must remain serial. And that serial part dominates most desktop applications, including games. That's why more cores don't help beyond two for the vast majority of desktop applications and games. See Amdahl's law: http://en.wikipedia.org/wiki/Amdahl%27s_law This is the limiting factor. And this is why I implore people to buy the fastest dual core and forgo the quad, six, eight core models. And in fact, for non gamer daily use, I recommend the single core AMD Sempron, because a dual core is wasted with Firefox, Thunderbird, Flash, Adobe Reader, etc. I failed to make a convincing case before you purchased Catherine. But at least you'll now be armed with this information when you make your next purchase. Financially it's not a huge deal, maybe $50 more in this case, 10% of the system price, for the quad core. But two cores will forever be wasted, and that $50 could have gone toward the discrete GPU you need. I am glad you finally understand that my desire for a dedicated video card with a substantial amount of dedicated video ram was not just a case of bigger, better, faster, more but based on the reality of what is required to do what I do. I can't imagine what screen resolution you thought I had in mind. 1024x768? It hurts my old eyes to look at those. We all make incorrect assumptions at times. I keyed on one thing you stated, which didn't get corrected until way late in the thread. That was I am retired. When I saw that I pegged you at 60+, or at least late 50s. I only know a few people in that age range, or older, playing WOW or any games, and they do it on big box brand PCs with integrated video on ~21 in screens. However, they run a resolution much lower than the panel native res so it's easier to read text without needing to use the 2nd/3rd lenses of their bi/trifocals, for instance 1280x720 instead of the panel native 1920x1080. At 1280x720 a discrete GPU is overkill for their needs. This was my mistake for not asking point blank early in the thread what res you were running instead of making assumptions based on your retired status. If I had asked more questions up front we could have avoided the contention. For that I apologize. I have also looked at my memory usage. At this very moment, not running WoW, I have 5.22 gig being used. 4 gig would not be sufficient for me. You would be correct if the number you're looking at reflected application memory usage. But it doesn't
Re: Building computer
On 10/2/2013 5:41 AM, Jeff Bauer wrote: On 10/02/2013 04:59 AM, Stan Hoeppner wrote: On 10/1/2013 5:13 PM, Catherine Gramze wrote: I have also looked at my memory usage. At this very moment, not running WoW, I have 5.22 gig being used. 4 gig would not be sufficient for me. You would be correct if the number you're looking at reflected application memory usage. But it doesn't. On any of the modern operating systems one must damn near be a computer scientist to see the actual memory usage. The 5.22GB, this is on Debian, yes? The system monitor? This reports process and cache memory usage. The buffer/cache will literally eat nearly all available memory all the time on Linux, then free some when an application process needs it. I've never used OSX but it's probably similar in its desktop reporting tool. This will really throw you for a loop. Open a shell window and execute ~$ sudo echo 3 /proc/sys/vm/drop_caches Wait a few seconds and see what happens to that 5.22GB number. Then report back what you find. You can do this while playing WOW as well. That number will drop like a rock and WOW will keep on going, because the memory you're freeing with that command is cache. And again, Linux will eat nearly all RAM for cache if the system is up long enough. free is another quick way to see where all your RAM went. Telling her to use free is premature at this point. Especially if you're not going to instruct her on how to use it. Otherwise it will only serve to confuse her. Which is precisely why I didn't mention it, at least not yet. You've apparently not been following the thread for the past week Jeff. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/524c16f8.9010...@hardwarefreak.com
Re: Getting Postfix to sort spam into different folders
On 10/2/2013 10:17 AM, Robin Kipp wrote:
Hi all,
I've been running my own mail server for a while (Postfix as the MTA and
Dovecot for retrieving mail via IMAP).
Yesterday, I added amavis-new for content filtering, and so far have
implemented spam checking using spamassassin and virus filtering using clamav.
All this appears to be working well - if anyone has some time to spare, feel
free to check the header of this message - if there are any things that could
be improved, I certainly would appreciate any feedback! :-)
So, now that spamassassin is flagging junk mail, I really would like a way of
having said junk sorted into a different folder for each user. I'm aware this
can also be done on the MUA side, but on some end-user devices (such as
smartphones), junkmail filtering is often not supported very unfortunately.
So, I was just wondering… Is there any way I could do this on the server side?
My main problem really is that I'm not sure where I'd best set this up - but
I'm guessing I'd have to do it in Postfix, as I use the 'virtual' local
delivery agent to drop incoming mail into the appropriate directories. Also,
the mailboxes are stored in maildir format if that makes any difference…
If anyone has anything like this working or could point me in the right
direction I would greatly appreciate that!
Many thanks :-)
Robin
Dovecot includes the sieve language for sorting. Per user, you'd have a
.dovecot.sieve file in the home directory, containing something like:
/home/stan/.dovecot.sieve
require fileinto;
if false {}
elsif header :contains X-Spam-Flag YES {
fileinto Spam;
stop;
}
...
This will move the flagged messages into the user's IMAP folder named
Spam.
You must use Dovecot LDA or LMTP for delivery if you want to use sieve.
The advantage of using sieve vs procmail, maildrop, etc, is that
Dovecot LDA/LMTP index the messages during the delivery phase. If you
use procmail/maildrop then Dovecot doesn't index messages until you open
the mailbox, which is slower.
You can also use Managesieve which allows users to create their own
filter rules. It is also possible to have a global sieve file for all
users, for thing such as spam sorting, and individual user defined sieve
rules simultaneously.
--
Stan
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/524ce3f0.1060...@hardwarefreak.com
Re: Building computer
On 10/2/2013 8:23 AM, Jonathan Dowland wrote: On Wed, Oct 02, 2013 at 03:59:15AM -0500, Stan Hoeppner wrote: This will really throw you for a loop. Open a shell window and execute ~$ sudo echo 3 /proc/sys/vm/drop_caches She'll probably get 'Permission denied', you need elevated permissions for the writing process, which in this case is the outer shell. Try instead $ sudo sh -c echo 3 /proc/sys/vm/drop_caches Thanks for the tip Jonathan. I never use sudo, so I was winging it a bit. And I wasn't about to instruct her to login as root or su to root... -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/524ce6d1.50...@hardwarefreak.com
Re: Building computer
On 10/2/2013 9:42 AM, Rhiamom wrote: On Oct 2, 2013, at 4:59 AM, Stan Hoeppner s...@hardwarefreak.com wrote This is the limiting factor. And this is why I implore people to buy the fastest dual core and forgo the quad, six, eight core models. And in fact, for non gamer daily use, I recommend the single core AMD Sempron, because a dual core is wasted with Firefox, Thunderbird, Flash, Adobe Reader, etc. I failed to make a convincing case before you purchased Catherine. But at least you'll now be armed with this information when you make your next purchase. Financially it's not a huge deal, maybe $50 more in this case, 10% of the system price, for the quad core. But two cores will forever be wasted, and that $50 could have gone toward the discrete GPU you need. I never suggested you were not correct about the CPU. I observed the lack of utilization of multiple cores on my first dual-core machine in 2006. I got The Haswell for the speed, lower power consumption, and presumably less heat generation. And possible resale value later on. This was my mistake for not asking point blank early in the thread what res you were running instead of making assumptions based on your retired status. If I had asked more questions up front we could have avoided the contention. For that I apologize. No apology needed. You did in fact peg my age correctly; I will be 61 next month. And I too have known people who run less than the native resolution to make the fonts bigger. When I get to that point, though, I will simply increase the font size so I don't get jaggies and blurry letters. Right now my eyes need the sharpness of the image, not a size increase. I'm lucky so far. In my early 40s and never needed glasses. Though I have noticed recently I must hold fine print further away to read it, such as product warning labels, the two paragraphs of exclusions on the phone/cable teaser rate flyers, etc. No issues with native res yet, thankfully. You would be correct if the number you're looking at reflected application memory usage. But it doesn't. On any of the modern operating systems one must damn near be a computer scientist to see the actual memory usage. The 5.22GB, this is on Debian, yes? The system monitor? This reports process and cache memory usage. The buffer/cache will literally eat nearly all available memory all the time on Linux, then free some when an application process needs it. I've never used OSX but it's probably similar in its desktop reporting tool. This was in OS X. The memory use would be similar in Debian, I assume. About A Wheezy desktop install isn't going to eat nearly as much RAM as desktop OSX, as there are far fewer service daemons loaded at startup. a quarter of the used memory was inactive which I assume was the cache. Still I'd guess that quarter is what has been paged to disk, not the cache. If you had no applications running, just the desktop and background processes, that shouldn't eat anywhere near 3.5GB. I'd think the cache is in what's left, not in the inactive. too close for comfort for me, as WoW was not running, nor ventrilo, and WoW does background downloads of the almost-weekly patches while you play, so even more processes. This will really throw you for a loop. Open a shell window and execute ~$ sudo echo 3 /proc/sys/vm/drop_caches Wait a few seconds and see what happens to that 5.22GB number. Then report back what you find. You can do this while playing WOW as well. That number will drop like a rock and WOW will keep on going, because the memory you're freeing with that command is cache. And again, Linux will eat nearly all RAM for cache if the system is up long enough. It is the nature of *nix to gobble up memory, yes. It will use what is there whether it needs it or not. Yes, the memory is used for block caching. The reason this is done is because the cost to drop cached lines is minimal when the memory is needed by an active process. So instead of wasting memory, Linux in particular, caches just about every block. But my Debian box has not arrived yet, so I can't run that command there. I could try it in terminal on my iMac, and it would probably work. This won't work on OSX. The /proc/sys/vm/ parameters in Linux are unique to Linux. I do thank you for the advice pertaining to a 384 bit bus and a gig more video ram than I was planning to get. That is advice that I will be following. You're welcome. Keep in mind that at 2560x1440 the 7950/7970 may still not be fast enough for full detail in WOW with GPU settings on high. The extra GB of VRAM won't get utilized but you need the memory bandwidth of a 384bit bus. Nobody sells, AFAICT, a 2GB model using these GPUs. I can't tell you where the setting resides, or if you have to edit xorg.conf, but you will want to use double buffering, not triple buffering. You'll also want to disable full screen
Re: Building computer
On 10/1/2013 12:29 AM, Rhiamom wrote: Sent from my iPad On Sep 30, 2013, at 10:33 PM, Stan Hoeppner s...@hardwarefreak.com wrote: ... It's quite funny to see someone of your knowledge level tell me I'm wrong by quoting the cardboard box as your evidence, while I'm demonstrating how the transistors and everything else work to get to a realistic set of requirements... This is the crux of the matter. Your requirements are not realistic for how Actually they were, up to the point you finally told us what screen resolution you use. That changes things quite a bit, or I should say changed one thing dramatically. I recommended a fast dual core CPU because more cores will be wasted. The i3-4340 3.6GHz Haswell would have been as fast as the quad core you bought, and saved ~$40-50. 4GB RAM, more than 4GB is wasted, but as I said previously it's cheap so buy more if you want, won't hurt. An SSD, and APU graphics. The only change in that recommendation, now that I know your screen resolution, is shifting to a very fast high bandwidth discrete card. 2560x1440 is a pretty insanely high gaming res if you want high frame rates and smooth rendering at high detail. Most hard core gamers wouldn't touch 2560x1440 without SLI/Xfire. If going w/a single reasonably priced card you're going to want/need a model with a 384 bit bus. Extremely high resolutions require extremely high memory bandwidth. The 384bit nVidia models are all above $600. The AMD 7950s can be had in the low $200s, and the 7970s in the low $300s. Both are 384bit. I want to use my computer. You may be able to happily exist on your minimal memory, ruthlessly eliminating background processes and OS features. I do not choose to do that. Your expert knowledge is worthless to me, because it requires me to alter the basic way I use my computer. In fact, it is worse than useless, because some poor sap might follow your advice and then wonder why they have performance issues with their brand new computer. No it doesn't change the way you use your computer. Because the specs I gave actually match how you currently use your computer. You simply don't know it, because you're not using the tools at your disposal which inform you of what system resources you're using. Run top, install Munin, etc, and look at the percentage of each CPU core that is used, and how much memory is used by your applications. You'll be very surprised. Then look at the GPU driver control panel while running WOW and see how much of the video RAM is in use. At 2560x1440 it may be pretty high. The 7950/7970 both sport 3GB of VRAM do you shouldn't fall short there. It doesn't matter as you already bought your system. But I find it interesting that you will be running integrated graphics for the time being, after you stated this is wholly inadequate. I also find it interesting that not once did you mention that you may try your old 6970 in the new box, before plunking down unnecessary cash on yet another high end video card. Yes, I will be running the integrated graphics for a few weeks while I adapt to the new box. It is only temporary. The 6970 is in my iMac, and will remain there. Note that even with 2 gig of dedicated video memory I am not able to play WoW on all high settings with the 6970. Again, that's not because there's not enough GPU memory, it's because the DRAM bus isn't fast enough, or the chip itself isn't fast enough, or both, for that insanely high resolution. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/524af3e9.5010...@hardwarefreak.com
Re: should an end user stick to a kernel with an initrd?
With scsi, the disk address is determined by its physical connection to the scsi cable. This is absolutely not correct. SCSI device IDs have always been programmed at the endpoint device via DIP switches, jumpers, or a dial. There was one short lived exception to this. In the late 1990s the industry created the SCAM extension, or SCSI Configured Automatically. This was an effort to make it easier for non technically savvy end users to install external devices such as scanners, optical drives, tape drives, without need to manually set device IDs. Cable proximity of devices was not involved in SCAM enumeration and assignment. It was a negotiated protocol. However, it was removed from the SCSI specification not long after introduction because it caused many more problems than it solved, specifically, it would change the boot hard drive ID and prevent systems from booting. It also wreaked havoc on software RAID systems when it auto reassigned IDs of the member drives, which obviously destroyed the array. I never fell victim to this because I didn't buy into the Adaptec hype. They started enabling SCAM by default in their controllers. I simply disabled it, and assigned IDs as I always had. A number of my colleagues did buy into the easy install hype and were bitten by it. -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/5249d3a8.2030...@hardwarefreak.com
