Re: munin related
In the last episode (Oct 05), Laszlo Danielisz said: Today while trying to install munin-node on 9.2 from ports I keep getting the following error: === Checking if sysutils/munin-common already installed === Creating users and/or groups. Using existing group `munin'. Creating user `munin' with uid `842'. pw: user 'munin' already exists *** [create-users-groups] Error code 74 Stop in /usr/ports/sysutils/munin-common. *** [build-depends] Error code 1 Stop in /usr/ports/sysutils/munin-node. *** [build] Error code 1 Stop in /usr/ports/sysutils/munin-node. Do you have any idea what can cause this? The ports are up to date. Do you have nscd caching enabled? It sometimes doesn't realize immediately that users/groups have been added to the system. Try restarting nscd, or disabling it temporarily while you install. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: How to ask a DNS resolver listening on a different port than the tcp/udp 53
On Sep 28, 2013, at 2:24 PM, Laurent SALIN salin.laur...@laposte.net wrote:
Le 28.09.2013 21:28, Mike. a écrit :
The way I solved this problem on my setup, I assigned another IP
address to the network interface via ifconfig alias.
I put the authoritative namesever on one IP address, and the
recursive nameserver on the other IP address.
They both are still listening on port 53, but on different IP
addresses.
hi,
If I could it would be just fine.
I got only one public IPv4 with each VPS. I've got a IPv6 too but I'm
not easy with IPv6 yet.
The provider (Tilaa) where I rent one of the 2 VPS, the one who may need
2 IPv4, is a bit short about his range of IPv4 and I guess it's not
raisonable to ask for a second IPv4 just for my personal use in case of
studying *BSD and networking stuff, I don't have a professional use here.
You only need to run one name server. It is both authoritative and recursive
by default. To limit recursion to only your own IP address space add the
following option in named.conf
options {
allow-recursion {
192.168.1.0/24;
127.0.0.1;
};
};
Change the address space to suit. Make sure you include localhost.
And after an rndc reload only your internal network will be able to make
recursive requests.
Dan
Thanks.
Laurent SALIN
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: test if script called by cron
In the last episode (Sep 16), Paul Macdonald said: Is there a simple way of testing whether a given script was called via cron, I'd rather find a solution that would work from within the script rather than setting an environment variable in the crontab. You check to see if stdin is a terminal, but that's not conclusive. One way to know for sure is to look at the name of the process that launched you: if [ ! -t 0 ] ; then echo no tty, possibly run from cron fi parent=$(ps -o command= -p $PPID) case $parent in *cron* ) echo parent is $parent, almost certainly cron ;; esac -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Geom Multipath
In the last episode (Sep 12), Outback Dingo said: does geom_multipath have some automatic type detection of mutipath drives? like in solaris? or is it all a manual process of labelling and such ?? It's all manual. Your two options are either gmultipath create, which will build a temporary device that will be forgotten on reboot, or gmultipath label, which writes a metadata block to the end of your disk. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: UUID in fstab.
In the last episode (Aug 26), John Baldwin said: On Monday, August 26, 2013 12:06:21 am varanasi sainath wrote: Thanks John, I have tried as you suggested using a Live CD and yes the partitions uuid's are present in gptid .. I found the UUID's in /dev/gptid - how do I determine which uid corresponds to which partition (ufs or swap or boot) (I used glabel status and after some trial and error I found them) edited the fstab accordingly and everything is working now .. gpart list will show detailed info for each provider, including the uuid for each GPT partition. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: ipfw confusion
You might turn on logging and post the logs of what is being blocked. Sometimes things are being blocked by rules you do not expect. On Mon, Aug 19, 2013 at 4:18 PM, Gary Aitken vagab...@blackfoot.net wrote: On 08/19/13 00:36, Jason Cox wrote: Are you sure that your DNS requests are over TCP? DNS primarily uses UDP to serve requests. TCP is used when the response data size exceeds 512 bytes (I think), or for tasks such as zone transfers. I know a few resolver implementations use TCP for all queries, but most I have used not. You might want to add rules to allow UDP as well. There are identical rules included for udp: 21149 allow udp from any to 12.32.44.142 dst-port 53 in via tun0 keep-state 21169 allow udp from any to 12.32.36.65 dst-port 53 in via tun0 keep-state One of the requests which is being refused is a zone transfer request from a secondary which is a tcp request. Others are probably udp. On Sun, Aug 18, 2013 at 11:06 PM, Gary Aitken vagab...@blackfoot.net wrote: I'm having some weird ipfw behavior, or it seems weird to me, and am looking for an explaination and then a way out. ipfw list ... 21109 allow tcp from any to 12.32.44.142 dst-port 53 in via tun0 setup keep-state 21129 allow tcp from any to 12.32.36.65 dst-port 53 in via tun0 setup keep-state ... 65534 deny log logamount 5 ip from any to any tail -f messages Aug 18 23:33:06 nightmare named[914]: client 188.231.152.46#63877: error sending response: permission denied 12.32.36.65 is the addr of the internal interface (xl0) on the firewall and is the public dns server. 12.32.44.142 is the addr of the external interface (tun0) which is bridged on a dsl line. It appears that a dns request was allowed in, but the response was not allowed back out. It seems to me the above rules 21109 and 21129 should have allowed the request in and the response back out. It's possible a request could come in on 12.32.44.142, which is why 21109 is present; although I know I am getting failures to reply to refresh requests from a secondary addressed to 12.32.36.65 What am I missing? Is there a problem if the incoming rule is for tun0, which gets passed to named since 12.32.44.142 is on the physical machine running named, but named pumps its response out on 12.32.36.65, relying on routing to get it to the right place, and that fails to match the state tracking mechanism which started with 12.32.44.142? ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: ipfw confusion
On Mon, Aug 19, 2013 at 1:06 AM, Gary Aitken vagab...@blackfoot.net wrote: ipfw list ... 21109 allow tcp from any to 12.32.44.142 dst-port 53 in via tun0 setup keep-state 21129 allow tcp from any to 12.32.36.65 dst-port 53 in via tun0 setup keep-state ... 65534 deny log logamount 5 ip from any to any What am I missing? Do you have a check-state rule earlier in your rules? 1000 check-state Dan ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Archiving a log file
In the last episode (Aug 04), Frank Leonhardt said: The answer isn't (AFAIK) newsyslog As a one-off, I need to archive an old log file - say httpd-access.log - while its still open. I don't want this to happen automatically and I don't want to set up newsyslog or anything like that. And I really don't want to mess about with signals to whatever is writing to the file, even assuming the writer could respond to them. I can't just rename the file as it's open for writing, and there would also be a good chance that something will be added to the file while it's being compressed. What I actually do is: cp httpd-access.log httpd-access.log-03-Aug-13 : httpd-access.log bzip2 httpd-access.log-03-Aug-13 Data might be lost here as something may be added between the cp being completed and the file being truncated. It's not the end of the world if this happens, but is there a better way? I could always shut down Apache for the duration, but I don't want to do that either, so in this case I'm happy to take the risk (it's not like I'm likely to miss anything that important). I don't know if this can be relied on as a POSIX thing, but the cp command simply(!) issues read() and write() calls until read() fails to get any more bytes, so if data is being appended to the file after cp is started it'll still be copied. Therefore the window where stuff could be written after the copy but before the truncation is shortened, but extant. So what's the magic utility I don't know about? newsyslog :) It renames the active logfile to a new name, sends the process a signal (syslog and SIGHUP by default) letting it know that it should close and reopen its logfile (creating a new one), then gzips the renamed file. In Apache's case, just specify the path to apache's pidfile in your newsyslog.conf, and everything should just work. Don't be afraid of signals. Without signalling the logging process to switch to a new logfile, and without suspending the process while you do your copy, there's always going to be a window where you risk losing logged data. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: TRIM on ZFS mirror
In the last episode (Aug 02), John Andreasson said: I read that FreeBSD 9.2 will bring TRIM to ZFS. Does anyone know if this works even if the zpool is a mirror? The vdev type doesn't matter. It'll work on plain disks, mirrors, and raidzs. If for some reason you have a pool built on geom mirrors, it'll work on that, too, since gmirror forwards BIO_DELETE requests to its components. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: TRIM and changing mount options
In the last episode (Jul 23), aurfalien said: On Jul 23, 2013, at 7:55 PM, Dan Nelson wrote: In the last episode (Jul 23), aurfalien said: At any rate, could some one; a) Explain how I am loading my file system as I'm used to fstab? b) How to run tunefs on my zroot c) How to determine if I actually have the needed TRIM support in my kernel I'm not sure if there's a way to query TRIM status on arbitrary geom providers, but you can see whether zfs successfully sent any TRIM requests by watching the output of sysctl kstat.zfs.misc.zio_trim. If .zio_trim.success increments and .unsupported (or .failed) doesn't, then you know that it's working. Looks like I don't have it. I keep reading that I must download the patch and make it which is not an issue. But I can't seem to find it. I've dl'd the 9.2 beta in hopes to check its src and see if its there. Looks like all I really need is the current rel. ZFS TRIM support was MFC'd into the 9 branch in June, so it wasn't in 9.1 but will be available in 9.2: http://svnweb.freebsd.org/base?view=revisionrevision=251419 -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: TRIM and changing mount options
In the last episode (Jul 23), aurfalien said: I've some what blindly followed a how to on installing FreeBSD 9.1 on a ZFS mirror. My typing is horrid so I simply ssh'd to a live CD env and pasted line by line as per the how to found here; http://wp.strahlert.net/wordpress/zfs-2/installing-freebsd-9-1-using-root-on-zfs-and-gpt-disks/ All is well, no issues with replacing disks, testing failures etc... But seeing that my system drives are SSDs, I thought to perhaps add noatime, etc... to avoid slow downs common in excessive SSD usage. However my fstab is only mounting swap partitions and I have no idea how my file system is being mounted :) Also, I would like to use tunefs in finding if I have TRIM enabled as I did load it via boot.conf I kinda feel like I'm willy nilly adding load lines w/o seeing if I actually have the TRIM patch installed :) At any rate, could some one; a) Explain how I am loading my file system as I'm used to fstab? b) How to run tunefs on my zroot c) How to determine if I actually have the needed TRIM support in my kernel ZFS automatically opens pools that were previously imported ( this info is stored in /boot/zfs/zpool.cache ), and automatically mounts any filesystems that have a mountpoint specified, so fstab isn't needed. Tunefs is a UFS-only command, so it won't help you here. Most zfs properties can be set on the fly with the zfs command. You can disable atime on your ZFS filesytem with zfs set atime=off zroot. All the filesystems created in zpool will inherit the value. I'm not sure if there's a way to query TRIM status on arbitrary geom providers, but you can see whether zfs successfully sent any TRIM requests by watching the output of sysctl kstat.zfs.misc.zio_trim. If .zio_trim.success increments and .unsupported (or .failed) doesn't, then you know that it's working. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: How to get file from nfs id
In the last episode (Jul 11), Radek Krejca said: Hello, I have problem with heavy load of my nfsd server. There is connected about 70 diskless machines, but in readonly mode. I catched traffic and get this: 21:00:39.715337 IP diskless-1.3297435097 storage.nfs: 112 getattr fh Unknown/A27801CEDE115FA30A005CD908007ABA4200570CDB51 21:00:39.716229 IP storage.nfs diskless-1.3297435097: reply ok 112 getattr REG 444 ids 0/0 sz 64944 21:00:39.716463 IP diskless-1.3297435098 storage.nfs: 112 getattr fh Unknown/A27801CEDE115FA30A0043DE08007CBA4200570CDB51 21:00:39.719112 IP storage.nfs diskless-1.3297435098: reply ok 112 getattr REG 444 ids 0/0 sz 82800 21:00:39.719453 IP diskless-1.3297435099 storage.nfs: 112 getattr fh Unknown/A27801CEDE115FA30A0043DE08007CBA4200570CDB51 21:00:39.721636 IP storage.nfs diskless-1.3297435099: reply ok 112 getattr REG 444 ids 0/0 sz 82800 Why is the same machine requesting chmod of the same file and is there any way to find out name of file from this long id? You can't get tcpdump to print it, but the getattr reply does include the inode of the file. If you start up wireshark and capture the same packets (or open a capture file created by tcpdump), the inode is stored in the nfs-obj_attributes-attributes-fileid field. You can then use find / -inum 12345 to locate that inode on disk. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Asus Sabertooth Z77 and FreeBSD?
Hello list Does anyone here have any experience with the Asus Sabertooth Z77 motherboard? How well does it work with FreeBSD and is all the hardware supported (including both SATA controllers)? Thanks! - Sincerely, Dan Naumov ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: showing CAPSLOCK state on display
In the last episode (Jun 20), Polytropon said: On Wed, 19 Jun 2013 21:19:01 +0200, Matthias Apitz wrote: I'm running 10-CURRENT on my small netbook EeePC 900; the device has only a set of four lights (powered on, battery loading, disk i/o, WLAN) and no indicator more, especially not for CAPSLOCK. So you don't know the state of it and have to try in in a terminal (which sometimes gives funny results when you write something with vim); some days ago I saw on a Windows 7 laptop that it showed CAPSLOCK in some small overlay text on-screen on the right sight. That would be just what I wanted for my KDE3 desktop... any ideas? Maybe something like this is suitable? Port: gai-leds-0.6_6 Path: /usr/ports/sysutils/gai-leds Info: A GAI applet that displays the keyboard status leds Port: gkleds2-0.8.2_6 Path: /usr/ports/sysutils/gkleds2 Info: GKrellM Leds for CapsLock, NumLock and ScrollLock But don't look at the dependency lists, they're terrible. :-) xkbvleds and xkbwatch might also do what you want, with no extra dependencies apart from X itself. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: apply /etc/ttys changes on system
In the last episode (Jun 19), takCoder said: Is there any way to apply /etc/ttys changes on system, except for *restarting system* and *running kill -HUP 1* command ? Due some reasons, i need to change tc value of some of my ttys, periodically. I'm looking for a safer way than *kill -HUP 1* command. So, as this command is not a good one to be used often, any ideas are really appreciated. I couldn't find any other one so far.. Why is kill -HUP 1 unsafe? It's documented in init's manpage: Line status (on, off, secure, getty, or window information) may be changed in the ttys(5) file without a reboot by sending the signal SIGHUP to init with the command ``kill -HUP 1''. On receipt of this signal, init re-reads the ttys(5) file. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: ZFS: raid VS copies=n
In the last episode (Jun 07), Quartz said: How does the ZFS option 'copies=n' and raid relate to and interact with each other? specifically recovery in the event of a failure. For example, is having three disks in a raid-1 configuration with copies=1 effectively the same as having three disks in a raid-0 with copies=3? Are the copies distributed uniformly across all drives in the pool, or concentrated, or what? What happens with configs like a raid-z2 with copies=2? Which / how many disks can you lose? The code will try to place the extra copies on different vdevs, but if that's not possible, it will try and place them at least 1/8th of the disk size apart on the same disk. Copies aren't meant to protect against whole disk loss, but more local damage within a disk. https://blogs.oracle.com/bill/entry/ditto_blocks_the_amazing_tape https://blogs.oracle.com/relling/entry/zfs_copies_and_data_protection -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: check variable content size in sh script
In the last episode (May 16), Tim Daneliuk said:
On 05/16/2013 10:08 AM, Joe wrote:
Hello
Have script that has max size on content in a variable.
How to code size less than 51 characters?
FOO=Some string you want to check length of
FOOLEN=`echo $FOO | wc | awk '{print $3}'`
You can then use $FOOLEN in a conditional.
Much better way:
FOO=Some string you want to check length of
FOOLEN=${#FOO}
--
Dan Nelson
dnel...@allantgroup.com
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: question on manpages/hier(7)
In the last episode (May 09), Paul Beard said: Where should site-specific, ie local, man pages live? For instance, I have: /usr/local/man/man1/php.1.gz /usr/local/share/man/man1/php.1.gz -rw-r--r-- 1 root wheel 3597 May 6 00:38 /usr/local/man/man1/php.1.gz -rw-r--r-- 1 root wheel 3383 Dec 20 19:54 /usr/local/share/man/man1/php.1.gz My understanding is that the older one is in the right place. The newer one is registered as belonging to php5.4-14 while the old one is orphaned. I don't have a /usr/local/share/man/ directory at all, and have 7300 files in /usr/local/man/man?/ , so I'd say /usr/local/man/ is the correct location :) -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Updating a broken link in page -http://adsm.vstyle.co.il/es/news/newsflash.html
Hello, I have bumped into your site while seeking for software and found your website very interesting :) Just a quick note, http://www.cdrom.com/is no longer active, and you are linking to it from page - http://adsm.vstyle.co.il/es/news/newsflash.html I was wondering if you don't mind updating the link to the updated websitehttp://en.downloadastro.com http://en.downloadastro.com/- A great download site with more than 300K software and games, including professional reviews and user ratings. We are now in a process of updating the links in websites all around the world and I will appreciate your cooperation. I am sure your *users will find it useful**.* Thanks, Dan ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
svnsync and local changes
Back when cvsup was in use, I mirrored the ports with cvsup-mirror. Then I could add files and make changes. My servers used my repository so they always had my changes and I only had to do them once. I am trying to replicate the same setup now that subversion is used. I've set up svnsync, and that works fine. I was able to add files with 'svn add' and 'svn commit'. My servers properly download all the new files. The problem is I can no longer use svnsync. Now it gives me Destination HEAD (316955) is not the last merged revision (316951).I've tried removing my files with 'svn rm', but that just changes the number in the error. Is there some way to get this to work with svnsync? Is there a better way for me to have a local repository that includes local changes? Thanks, Dan ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Copying memstick image to a USB (flash/thumb) drive
In the last episode (Mar 27), Ronald F. Guilmette said: I've never used any FreeBSD memstick image before, but now I have reason to do so. I'm reading the instructions for creating a bootable memstick that are located on this page: http://www.freebsd.org/releases/9.1R/announce.html which include the following example of how to perform the copy: # dd if=FreeBSD-9.1-RELEASE-amd64-memstick.img of=/dev/da0 bs=10240 conv=sync Question: Why exactly is conv=sync is there? Question: Why exactly is the bs=10240 is there? Wouldn't the default of 512 do just as well? It looks like someone just copied a dd commandline from somewhere else, maybe something to do with tar files (since tar defaults to a 10k blocksize when writing to tape). conv=sync isn't needed since the source file is already a multiple of the target device blocksize (512 bytes), and bs=64k would be much faster when writing to cheap flash devices like USB sticks since they don't have a write cache and individual writes are slowish. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Leaking disk space
A stab in the dark, but does # sync change anything Alas, no. On 21 March 2013 13:21, Bernt Hansson b...@bananmonarki.se wrote: On 2013-03-21 11:40, Dan Thomas wrote: Have you used fstat to identify the big growing file which is taking up the space, and which process has the file open? It's not an unlinked file. I've tried using fstat and lsof to identify it, and there's no inodes with zero links or that don't have a matching file on disk. A stab in the dark, but does # sync change anything. Dan On 20 March 2013 23:08, Daniel O'Callaghan da...@clari.net.au wrote: On 21/03/2013 3:55 AM, Dan Thomas wrote: Stopping Postgres doesn't fix it, but rebooting does which points at Have you used fstat to identify the big growing file which is taking up the space, and which process has the file open? A file which has been unlinked from all directories won't be seen by du, but it does not free disk space until no process has it open. USER CMD PID FD MOUNT INUM MODE SZ|DV R/W root syslogd476488 /4317027 -rw-r--r-- 19776 w root syslogd476489 /4317041 -rw--- 63 w That might help to track it down. Danny ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Leaking disk space
Have you used fstat to identify the big growing file which is taking up the space, and which process has the file open? It's not an unlinked file. I've tried using fstat and lsof to identify it, and there's no inodes with zero links or that don't have a matching file on disk. Dan On 20 March 2013 23:08, Daniel O'Callaghan da...@clari.net.au wrote: On 21/03/2013 3:55 AM, Dan Thomas wrote: Stopping Postgres doesn't fix it, but rebooting does which points at Have you used fstat to identify the big growing file which is taking up the space, and which process has the file open? A file which has been unlinked from all directories won't be seen by du, but it does not free disk space until no process has it open. USER CMD PID FD MOUNT INUM MODE SZ|DV R/W root syslogd476488 /4317027 -rw-r--r-- 19776 w root syslogd476489 /4317041 -rw--- 63 w That might help to track it down. Danny ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Leaking disk space
Hi Guys,
We're seeing a problem with some of our FreeBSD/PostgreSQL servers
leaking quite significant amounts of disk space:
df -h /usr/local/pgsql/
Filesystem SizeUsed Avail Capacity Mounted on
/dev/mfid1s1d1.1T772G222G78%/usr/local/pgsql
du -sh /usr/local/pgsql/
741G/usr/local/pgsql/
Stopping Postgres doesn't fix it, but rebooting does which points at
the OS rather than PG to me. However, the leak is only apparent in the
dedicated pgsql partition, and only on our database servers, so
PostgreSQL seems to at least be involved. The partition itself is a
relatively standard UFS partition:
grep /usr/local/pgsql /etc/fstab
/dev/mfid1s1d /usr/local/pgsqlufs rw 2 2
tunefs -p /usr/local/pgsql/
tunefs: POSIX.1e ACLs: (-a)disabled
tunefs: NFSv4 ACLs: (-N) disabled
tunefs: MAC multilabel: (-l) disabled
tunefs: soft updates: (-n) enabled
tunefs: gjournal: (-J) disabled
tunefs: trim: (-t) disabled
tunefs: maximum blocks per file in a cylinder group: (-e) 2048
tunefs: average file size: (-f)16384
tunefs: average number of files in a directory: (-s) 64
tunefs: minimum percentage of free space: (-m) 8%
tunefs: optimization preference: (-o) time
tunefs: volume label: (-L)
LSOF isn't showing any open files:
lsof +L /usr/local/pgsql/ | awk '{ print $8 }' | grep 0 | wc -l
0
We're not creating filesystem snapshots:
find /usr/local/pgsql/ -flags snapshot
Not all of our servers are leaking space, it's only the more
recently-installed systems. Here's a quick breakdown of versions:
FreeBSD PostgreSQL Leaking?
8.0 8.4.4no
8.2 9.0.4no
8.3 9.1.4yes
8.3 9.2.3yes
9.1 9.2.3yes
Any ideas what's going on here, or where we could start debugging?
Thanks,
Dan
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Leaking disk space
a) Where do you have the wal files?
pg_xlog is symlinked to /usr/local/pglog/pg_xlog (ie, out of the
partition mounted as /usr/local/pgsql which is exhibiting this
behaviour).
b) Are you sure that unused/old wal files are erased?
As above, but yes they seem to be being deleted properly
c) Do you have any postgres log level activated (like the ones used
for long queries)?
Yes we have slow query logging enabled. pg_log is symlinked out of
that partition to /usr/local/pglog/pg_log as well.
d) Does your queries have GROUP BY on very big data sets? Those create
big temporal data files.
Yes we do a lot of that! However there are definitely no unlinked
files, and the problem doesn't go away when pg is shut down. However a
reboot does fix it.
e) With question a) and b), do you use streaming replication?
Yes we do. This problem is not present on the warm standby servers
that are being streamed to. We have failed over to the warm standbys
previously (we're currently doing this regularly to work around the
problem without too much downtime). Once we switch the warm standby to
primary, it begins leaking space.
On 20 March 2013 16:02, Eduardo Morras emorr...@yahoo.es wrote:
On Wed, 20 Mar 2013 15:23:18 +
Dan Thomas godd...@gmail.com wrote:
Hi Guys,
We're seeing a problem with some of our FreeBSD/PostgreSQL servers
leaking quite significant amounts of disk space:
df -h /usr/local/pgsql/
Filesystem SizeUsed Avail Capacity Mounted on
/dev/mfid1s1d1.1T772G222G78%/usr/local/pgsql
du -sh /usr/local/pgsql/
741G/usr/local/pgsql/
Stopping Postgres doesn't fix it, but rebooting does which points at
the OS rather than PG to me. However, the leak is only apparent in the
dedicated pgsql partition, and only on our database servers, so
PostgreSQL seems to at least be involved. The partition itself is a
relatively standard UFS partition:
grep /usr/local/pgsql /etc/fstab
/dev/mfid1s1d /usr/local/pgsqlufs rw 2 2
tunefs -p /usr/local/pgsql/
tunefs: POSIX.1e ACLs: (-a)disabled
tunefs: NFSv4 ACLs: (-N) disabled
tunefs: MAC multilabel: (-l) disabled
tunefs: soft updates: (-n) enabled
tunefs: gjournal: (-J) disabled
tunefs: trim: (-t) disabled
tunefs: maximum blocks per file in a cylinder group: (-e) 2048
tunefs: average file size: (-f)16384
tunefs: average number of files in a directory: (-s) 64
tunefs: minimum percentage of free space: (-m) 8%
tunefs: optimization preference: (-o) time
tunefs: volume label: (-L)
LSOF isn't showing any open files:
lsof +L /usr/local/pgsql/ | awk '{ print $8 }' | grep 0 | wc -l
0
We're not creating filesystem snapshots:
find /usr/local/pgsql/ -flags snapshot
Not all of our servers are leaking space, it's only the more
recently-installed systems. Here's a quick breakdown of versions:
FreeBSD PostgreSQL Leaking?
8.0 8.4.4no
8.2 9.0.4no
8.3 9.1.4yes
8.3 9.2.3yes
9.1 9.2.3yes
Any ideas what's going on here, or where we could start debugging?
Somethings to check:
a) Where do you have the wal files?
b) Are you sure that unused/old wal files are erased?
c) Do you have any postgres log level activated (like the ones used for long
queries)?
d) Does your queries have GROUP BY on very big data sets? Those create big
temporal data files.
e) With question a) and b), do you use streaming replication?
Thanks,
Dan
--- ---
Eduardo Morras emorr...@yahoo.es
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Upgrade from 6.4 to 9.1?
In the last episode (Mar 16), Drew Tomlinson said: I have an old system happily running 6.4 but am finding that it can no longer download or build a ports index. Thus I guess it's time to upgrade. What gotchas do I need to look out for? In the past my upgrades have always been simply downloading new source, reviewing kernel config file, and then rebuilding the system. Any ports that didn't work after that I would just rebuild as well. However I've never waited this long to upgrade. Do I need to do anything different? You won't be able to do a straight source build from 6.4 to 9.1; too many low-level changes like Makefile syntax and compiler options have changed. If you are comfortable with temporarily disabling non-essential things that fail to build, it is definitely possible to do a long jump to 9.1, but it'd be safer to either hop from 6.4 - (7-stable or 8-stable) - 9 doing buildkernels and buildworlds, or just do a binary upgrade of kernel and base system to 9.1. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: NTFS 4 Terabyte drive problems on FreeBSD?
In the last episode (Feb 12), Scott Ballantyne said: Hi, I'm trying to mount a 4 TB drive (The Seagate Backup-Plus) on FreeBSD 9.0 Generic, using mount -t ntfs /dev/da0s1 /mnt It fails with: g_vfs_done() da1s1[READ(offset=0, length=1042)]error=22 mount_ntfs: /dev/da1s1 bad argument That's a strange length (1042). I would have expected a multiple of the device blocksize (4096 in the case of the Backup Plus). However, I can mount a 2 TB ntfs drive with no problem, using the same command, and the 4TB drive checks out fine in Windows. Does anyone know if this limitation exists, and if there is a work-around? I only need READ access to this drive. You can try the sysutils/fusefs-ntfs port; the ntfs code in the base system hasn't had much work done on it lately. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: sh script code to get file size.
In the last episode (Jan 18), Fbsd8 said: In a script in am working on I need to find out the allocated size of a sparse file. The only command that comes to mind is ls -lh The du -h command is not appropriate because it will show the occupied size and not the allocated size. I don't know how to parse out to the position in the output of that ls -lh command to pickup the file size value. Is there some other way to do this? For a single file, you can get the exact filesize with stat -f %z filename. For the human-friendly version, du -hA will work. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: OT: What Might Break getbostbyname() ?
In the last episode (Jan 16), Tim Daneliuk said: This is not really a FreeBSD problem ... in fact, it's happening on a Solaris 10 machine. But because the TCP stack and its userland interface came from BSD, I am hoping some kind soul might have an insight into what's going on ... Solaris hasn't used a BSD TCP stack for many many years afaik.. The machine in question does DNS lookups fine via dig or nslookup. I believe these connect directly to the DNS server(s) specified in /etc/resolv.conf. However, any program that uses gethostbyname() - like ping - fails and says it cannot resolve the name. I'm looking for hints here on why or how gethostbyname() and/or the network stack could get clobbered so as to not be able to talk to the DNS servers which I know are reachable via dig and nslookup. First, check /etc/nsswitch.conf and verify that dns is listed on the hosts: line. Next, try disabling nscd (svcadm disable name-service-cache) , and then running truss ping www.google.com (make sure to reenable nscd when you're done debugging). You should see syscalls to open /etc/resolv.conf, read the contents, and then open a socket to the nameserver listed in that file. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: assigning values to variables in the background
In the last episode (Jan 16), Nikos Vassiliadis said: On 1/15/2013 9:30 PM, Greg Larkin wrote: On 1/15/13 12:42 PM, Nikos Vassiliadis wrote: A bit of an OT question. I am writing a bourne shell script that collects data from a router. Since netstat vmstat can run for a numbers of iterations I thought I would use just that: stats() ( nstats=`netstat -I ng0 -q 1 60 | tail -1` rawdata=`vmstat -c 2 60 | tail -1` wait ... The logic was: 1. run the first process in the bg 2. run the second process 3. wait to make sure the first process has finished 4. continue further It makes perfect sense why this doesn't work. Both commands run in the foreground. I am going to split the time between netstat and vmstat. So, it will be 30 seconds of netstat and 30 seconds of vmstat. But I am still interested/curious how one should go for this using the shell. So, can this be done without files? Any thoughts? As far as I can tell, the backticks are what's causing the problem. Even though you put the first command in the background (maybe with the inside the backticks, though), the assignment to the nstats variables causes the script to block. If you switch to using temp files, you may have more luck, e.g.: netstat -w 1 -I ng0 -q 60 | tail -1 /tmp/netstat.$$ npid=`echo $!` vmstat -w 2 -c 60 | tail -1 /tmp/vmstat.$$ vpid=`echo $!` wait $npid nstats=`cat /tmp/netstat.$$` rm -f /tmp/netstat.$$ wait $vpid rawdata=`cat /tmp/vmstat.$$` rm -f /tmp/vmstat.$$` npid=$! is cleaner (no need to fork a subshell just to echo a variable), but you don't even need that. You can use just a single wait command to wait for both processes to finish, then extract the output of both tempfiles. Yes, this looks probably like something I will use too. Just for the fun of it and using a separator(Robert's idea), I came up with this: delay=10 a=$( ( echo netstat `netstat -I ng0 -q 1 $delay | tail -1` netstat ) ( echo vmstat `vmstat -c 2 $delay | tail -1` vmstat ) ) echo $a This works, but now you have both lines of info in a single variable, and it may be more work to split the lines back out (also note that you can't predict which line will be first). If you want to stick with shell, either zsh or bash would make short work of parsing this. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: update to SASL breaks sendmail
In the last episode (Jan 05), Robert Huff said: I have followed the canonical procedure to get Sendmail to use SASL. Yesterday this worked. This morning I updated cyrus-sasl to the latest version, which bumps the library version from .2 to ,3. This appears to break sendmail in at least two places. I have added a mapping in libmap.conf ... which seems to work ... but I'm pretty sure that's Not The Right Thing. What is? Rebuild sendmail so that it links against the updated sasl libraries, or make sure to preserve old shared libraries when upgrading packages. This isn't sendmail or sasl-specific; it's standard procedure when you upgrade any port that installs shlibs that other ports depend on. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: update to SASL breaks sendmail
In the last episode (Jan 05), Robert Huff said: On 1/5/2013 8:30 PM, Dan Nelson wrote: In the last episode (Jan 05), Robert Huff said: This morning I updated cyrus-sasl to the latest version, which bumps the library version from .2 to ,3. This appears to break sendmail in at least two places. I have added a mapping in libmap.conf ... which seems to work ... but I'm pretty sure that's Not The Right Thing. What is? Rebuild sendmail so that it links against the updated sasl libraries, or make sure to preserve old shared libraries when upgrading packages. I know how to build world; what is the correct way of building/installing just sendmail and making sure I get the right libraries? (The information in /usr/ports/security/cyrus-sasl/files/Sendmail.readme doesn't appear to apply.) Base sendmail doesn't link with sasl by default. If you had edited Makefiles or make.conf to enable that, running make clean all install clean in /usr/src/usr.sbin/sendmail/ should build and install just the new sendmail. Or, if you had installed the mail/sendmail port with sasl enabled, force-upgrade that port using your favorite method. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Dialog on some ports looks odd
Hey there, Can people confirm some brokenness to me? When I'm on a system over SSH, I find that doing the following: cd /usr/ports/mail/alpine; make config looks fine, but cd /usr/ports/mail/opendkim; make config seems to corrupt the headings and not display correctly, the OK/Cancel buttons get mangled (it may or may not work on the system console). Could I get some confirmation before I do a send-pr? -- I can feel it, comin' back again...Like a rolling thunder chasin' the wind... -Dan Mahoney, JS, JB SL, May 10th, 1997, Approx 1AM Dan Mahoney Techie, Sysadmin, WebGeek Gushi on efnet/undernet IRC ICQ: 13735144 AIM: LarpGM Site: http://www.gushi.org --- ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: what replaces javaws? using icedtea-web and openjdk6.
In the last episode (Dec 06), Antonio Olivares said: http://www.freebsd.org/cgi/query-pr.cgi?pr=ports/173603 I apply the suggested fix: $ sh -x `which itweb-javaws` jviewer.jnlp + JAVA=/usr/local/openjdk6/jre/bin/java + LAUNCHER_BOOTCLASSPATH=-Xbootclasspath/a:/usr/local/share/icedtea-web/netx.jar + LAUNCHER_FLAGS=-Xms8m + CLASSNAME=net.sourceforge.jnlp.runtime.Boot + BINARY_LOCATION=/usr/local/bin/itweb-javaws + PROGRAM_NAME=itweb-javaws + CP=/usr/local/openjdk6/jre/lib/rt.jar /usr/local/bin/itweb-javaws: 11: Syntax error: Bad function name I try once more on another machine not 64 bit, it returns the same error and java web start does not work :( $ sh -x `which itweb-javaws` jviewer.jnlp + JAVA=/usr/local/openjdk6/jre/bin/java + LAUNCHER_BOOTCLASSPATH=-Xbootclasspath/a:/usr/local/share/icedtea-web/netx.jar + LAUNCHER_FLAGS=-Xms8m + CLASSNAME=net.sourceforge.jnlp.runtime.Boot + BINARY_LOCATION=/usr/local/bin/itweb-javaws + PROGRAM_NAME=itweb-javaws + CP=/usr/local/openjdk6/jre/lib/rt.jar /usr/local/bin/itweb-javaws: 11: Syntax error: Bad function name Any other ideas as to how to fix this? Don't try and run it through /bin/sh . The script uses bash-isms (array syntax specifically). Just run itweb-javaws jviewer.jnlp. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: getpwnam_r returns EINVAL on FreeBSD 8.3
On Mon, Dec 3, 2012 at 5:54 PM, Dan Lists lists@gmail.com wrote:
After upgrading a server from FreeBSD 7.3 to FreeBSD 8.3 I noticed
this bug. Since upgrading, getpwnam_r is acting inconsistently. If I
look up a user that does not exist and the name is 16 characters or
less, getpwnam_r returns 0 and the result is NULL. If the name is
more than 16 characters, getpwnam_r returns EINVAL. Everything works
correctly for users that exist.
I was incorrect. The behavior of getpwnam_r is the same on 7.3 and
8.3. The software I was testing acted differently on the two
versions.
This only happens when the nsswitch.conf passwd: line contains files.
You need to use files if you are using another module such as msql or
ldap. The problem exists without the other modules listed. For
example:
passwd: files
I would like to emphasize that this does NOT happen when passwd: is
set to compat. I believe this is a bug. getpwnam_r should have the
same return values (or errno for getpwnam) whether nsswitch.conf has
compat or files. If there is a really good reason for them to be
different, it should be documented.
Below is a simple test program. Set passwd: to files in nsswitch.conf
and run the program. Any idea how to fix this bug with getpwnam_r?
#include stdio.h
#include sys/types.h
#include pwd.h
main()
{
lookup(doesnotexist);
lookup(doesnotexisty);
}
int lookup( char *name)
{
struct passwd pwd;
char buffer[1024];
struct passwd *result;
int err;
printf(\nLooking up: %s\n, name);
err = getpwnam_r(name, pwd, buffer, sizeof(buffer), result);
if( err != 0 ){
printf(Return code: %d\n, err);
}else if( result == 0 ){
printf(Returned no result!\n);
}else{
printf(Returned: %s (%d)\n, result-pw_name, result-pw_uid);
}
}
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: redirect incoming telnet to com port
In the last episode (Dec 03), saeedeh motlagh said: hello everybody i have freebsd8.2. i want to redirect incoming telnet from telnet port to a com port. i mean if somebody telnet to my system with a specific port (or 22 which is default telnet port), connect to com port and can talk to modem (something like cisco). (Port 23 is telnet, btw. Port 22 is ssh) is it possible or not? There are probably a dozen ports that do this, all in the comms category of the ports tree. You can choose anything from a dumb telnet interface with no line control (comserv, ser2net, tcpser), to a RFC-2217 interface that gives you full low-level control over a remote serial port (sredird), to a managed system like with optional encryption and per-user access control (conserver). -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
getpwnam_r returns EINVAL on FreeBSD 8.3
After upgrading a server from FreeBSD 7.3 to FreeBSD 8.3 I noticed
this bug. Since upgrading, getpwnam_r is acting inconsistently. If I
look up a user that does not exist and the name is 16 characters or
less, getpwnam_r returns 0 and the result is NULL. If the name is
more than 16 characters, getpwnam_r returns EINVAL. Everything works
correctly for users that exist.
This only happens when the nsswitch.conf passwd: line contains files.
You need to use files if you are using another module such as msql or
ldap. The problem exists without the other modules listed. For
example:
passwd: files
Below is a simple test program. Set passwd: to files in nsswitch.conf
and run the program. Any idea how to fix this bug with getpwnam_r?
#include stdio.h
#include sys/types.h
#include pwd.h
main()
{
lookup(doesnotexist);
lookup(doesnotexisty);
}
int lookup( char *name)
{
struct passwd pwd;
char buffer[1024];
struct passwd *result;
int err;
printf(\nLooking up: %s\n, name);
err = getpwnam_r(name, pwd, buffer, sizeof(buffer), result);
if( err != 0 ){
printf(Return code: %d\n, err);
}else if( result == 0 ){
printf(Returned no result!\n);
}else{
printf(Returned: %s (%d)\n, result-pw_name, result-pw_uid);
}
}
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: SSD for FreeBSD NAS device
In the last episode (Nov 20), bsd said: I have just acquired an Intel R2312GZ4GC4 which I have equipped with a Adaptec RAID 51245 and 6 WD red disks of 3To - It'll come with 32Gb of Kingston ECC RAM. I am planing to use It as a backup device on a second hosting facility to backup couple of critical servers of mine. [..] I wanted to know what were your experiences on choosing an SSD HD as master boot device / root FS ? Do you think I should go for a redundant SSD drives (RAID 1) or does this offers limited interest in such config ? For any critical server, don't think of RAID as an option, think of it as a requirement. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Issues with smartd starting up at boot time - delays sever start?
In the last episode (Nov 13), Karl Pielorz said: I've noticed on our systems (9.0-Stable, amd64) that starting smartd at boot time massively extends the startup time of the box. I think I've traced this down to smartd, and our use of the '-M test' config option (which sends a test message, apparently forking to 'mail' - and, as the config man page says - it will block until that command returns). For whatever reason (networking not stable at that point in time, MTA not started yet etc.) - on our machines this leaves smartd handing around for minutes - before it returns, the machine starts up (and the status emails arrive). Can anyone think of a 'simple' fix for this? - Is there anything I can do to '/usr/local/etc/rc.d/smartd' to make it run later in the startup process? Try adding mail to the REQUIRE: line, since sendmail has that in its PROVIDES: line. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Linker Flags against Kerberos5 Libraries
Hey all, I've been working with someone doing development against the stock kerberos libraries installed in FreeBSD 8.3, and they've noticed that the symbols present in some of the kerberos libraries don't stand on their own. For example: configure:13089: cc -o conftest -O2 -pipe -fno-strict-aliasing -I/usr/local/include -L/usr/local/lib conftest.c -lkrb5 -lasn1 -lcom_err -lcrypto -lcrypt -lpam 5 /usr/lib/libasn1.so: undefined reference to `rk_strpoolcollect' /usr/lib/libkrb5.so: undefined reference to `bswap16' /usr/lib/libkrb5.so: undefined reference to `hx509_certs_init' /usr/lib/libkrb5.so: undefined reference to `hx509_cert_get_serialnumber' /usr/lib/libkrb5.so: undefined reference to `strlwr' /usr/lib/libasn1.so: undefined reference to `rk_hex_encode' /usr/lib/libkrb5.so: undefined reference to `hx509_cert_find_subjectAltName_otherName' /usr/lib/libkrb5.so: undefined reference to `hx509_certs_find' /usr/lib/libkrb5.so: undefined reference to `net_write' /usr/lib/libkrb5.so: undefined reference to `socket_sockaddr_size' /usr/lib/libkrb5.so: undefined reference to `hx509_verify_attach_anchors' [...] From this, it would seem to be the case that any program linked against any of the Kerberos libraries will fail unless linked against ALL of them, and the libraries don't seem to depend on each other. My questions: 1) Is there a sane way of detecting the required linker flags? 2) Is there a better mailing list I should be asking these on? -Dan ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: harware compatibility queustion
In the last episode (Aug 23), Robert Bonomi said: I need to add more than two' internal SATA _ports_ to a box that will be running FreeBSD 9. When I look at inexpensive SATA controllers, they all seem to be using chipsets thar are _not_ explicitly mentioned in the 'supported hardware' list. Can anyone say, authoritatively, that, say, a 'SiL3114'-based cnntroller card will work in FreeBSD, as far as 'bare drive' functionality goes? The ata manpage says it supports the SiI3114 card (even the siliconimage.com webpage uses both SiI and Sil spellings all over), so you should be safe. Alternatively, can anybody recommend an inexpensive 4-port (or more) SATA controller card that is known to work with FreeBSD? -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: firefox png
In the last episode (Aug 22), david coder said: thx, i hadn't seen the reply to my earlier message. unfortunately, though i've got the png port installed w/ OPTIONS=APNG Animated PNG support On This line just tells you what the default is on a system that hasn't built the png port yet. Before 2011-07-08, the default was Off. the install of freebsd fails w/ the error message given below. What does make showconfig in the ports/graphics/png directory print? I bet you first installed the png port over a year ago, so you have inherited the Off value from then. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Disk Errors
In the last episode (Jul 24), dweimer said: I have three 1TB disks I use for backup, two of them are Western Digital drives I bought specifically for this purpose. One is a Seagate drive that came out of a barebones PC that I replaced with a couple smaller drives in a stripe to gain performance. I use the drives in an external SATA dock, using geom eli encryption, the western digital drives give me no problems, but the seagate drive gives me a lot of the following errors under load. ad4: TIMEOUT - WRITE_DMA48 retrying (1 retry left) LBA=817755328 ad4: WARNING - WRITE_DMA48 UDMA ICRC error (retrying request) LBA=837397120 ad4: TIMEOUT - WRITE_DMA48 retrying (1 retry left) LBA=879786112 ad4: WARNING - WRITE_DMA48 UDMA ICRC error (retrying request) LBA=882931200 ad4: WARNING - WRITE_DMA48 UDMA ICRC error (retrying request) LBA=890542016 ad4: WARNING - WRITE_DMA48 UDMA ICRC error (retrying request) LBA=902767296 ad4: TIMEOUT - WRITE_DMA48 retrying (1 retry left) LBA=904071296 If you install the sysutils/smartmontools port, you can run smartctl -x /dev/ad4 to dump the drive's SMART attribute table and error logs. Those should give you an indication of whether the drive is going bad. If the drive is logging those write errors in its internal log, then you know it's not a cabling issue. If it's not logging errors, I suppose you might have a loose SATA plug on the drive itself, which would explain why the problem follows the drive around. dmesg info about the drive at connection time: ad4: 953869MB Seagate ST31000528AS CC46 at ata2-master UDMA100 SATA 3Gb/s dmesg info about one of the western digital drives: ad4: 953869MB WDC WD10EARS-00Y5B1 80.00A80 at ata2-master UDMA100 SATA 3Gb/s Before I scrap the drive I just wanted to see if anyone could either say for sure its hardware, or if something else could possibly cause this. I don't suspect the controller, cable or dock as the problems would likely occur with the western Digital drives as well if one of them were involved. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
portupgrade -- is there a way to only build and update ports that actually NEED it?
Hey there, I'm presently in the process of trying to do a portupgrade from rt-3.8.8 to 3.8.13. By all estimations, this is a minor bump. Already, I've encountered several annoyances due to ABI changes, such as the libtool2.4 fun. With normal portupgrade, this forces you to go fix the dependent port. Finally, I just applied -r, which should update all dependent packages, but it seems to upgrade them unconditionally. Ergo, I've since built a new version of perl, a new verion of python, rebuilt every perl module on the system, am presently rebuilding apache22, and I'm sure the system will turn around and require me to rebuild postgres real soon. You would think there's an option to portupgrade that says don't upgrade every single package I've got, but if somewhere in the dependency chain I need a newer version of a thing, then do it. Am I just missing it in the manpages, or does such a thing really not exist? -Dan -- You recreate the stars in the sky with cows? -Furrball, March 7 2005, on Katamari Damacy Dan Mahoney Techie, Sysadmin, WebGeek Gushi on efnet/undernet IRC ICQ: 13735144 AIM: LarpGM Site: http://www.gushi.org --- ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Newbie question: Why aren't my cron jobs running?
On Tue, Jun 12, 2012 at 12:06 PM, Polytropon free...@edvax.de wrote: On Tue, 12 Jun 2012 08:29:02 -0500, Mark Felder wrote: On Tue, 12 Jun 2012 00:06:21 -0500, Robert Bonomi bon...@mail.r-bonomi.com wrote: Comment: using a leading zero on the numeric fields is a BAD IDEA(tm) -- you are *strongly* encocuraged to remove them. Yes, that means numbers will not be column aligned, but it is a small price to pay to avoid the hair-tearing that =will= ensue when using it bites you. Any other info on this? I've never heard of this before and I've never seen an issue using leading zeroes on the minutes value. There are some specific interpretations that _may_ be interpreted according to the C rules, e. g. prefix 0x- for hexadecimal or 08- for octal notation. For example, 083 != 83, just as 0x83 != 83. As it has been mentioned, spaces also have a significant meaning in crontabs, so they cannot be used everywhere to align data columns. The syntax of his crontab file is correct. Vixie cron does care about leading spaces, tabs, extra spaces, or leading zeros. Earlier versions of cron are much pickier about the crontab file. The cron logs show that it is starting his jobs at the correct times. It is far more likely that there is a problem with the scripts. A very common cause of problems with scripts run from cron is that they do not inherit your environment. Do the scripts run from the command line? If the do, then the problem is most likely something in your environment that the scripts need. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
possbility of a port for older versions of libintl?
Hey there, I recently discovered that the vmware-tools package is compiled against libintl.so.8 -- yes, this is probably something that should be fixed at the vmware level, but VMware's love for FreeBSD isn't there. As a workaround, it might be useful to have a port which compiles an older version of libintl (potential security issues notwithstanding, since it's assumed it will only be used by this one tool). This seems to me to be somewhat *less* destabilizing than the commonly-suggested (and perhaps, oft-used) suggestion of symlinking /usr/lib/libintl.so.8 -- libintl.so.9 Thoughts? -- Dan Mahoney Techie, Sysadmin, WebGeek Gushi on efnet/undernet IRC ICQ: 13735144 AIM: LarpGM Site: http://www.gushi.org --- ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Proper Port Forwarding
In the last episode (Jun 06), Michael Sierchio said: On Wed, Jun 6, 2012 at 11:31 AM, Simon si...@optinet.com wrote: This easily causes DoS for when too many FIN_WAIT_2 are created and IPFW stops forwarding using the rule above because of too many dynamic rules Change the defaults for the fw.dyn sysctl MIB nodes to something like net.inet.ip.fw.dyn_short_lifetime=3 net.inet.ip.fw.dyn_udp_lifetime=3 net.inet.ip.fw.dyn_rst_lifetime=1 net.inet.ip.fw.dyn_fin_lifetime=1 net.inet.ip.fw.dyn_syn_lifetime=10 Or raise net.inet.ip.fw.dyn_max to a larger number. The default 4096 may be too small. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Possible /bin/sh Bug?
In the last episode (Jun 05), Tim Daneliuk said: Given this script: #!/bin/sh foo= while read line do foo=$foo -e done echo $foo Say I respond 3 times, I'd expect to see: -e -e -e Instead, I get: -e -e Linux appears to do the right thing here, so this seems like it is a bug ... or am I missing something? echo takes a -e flag, so it eats the first one. Bash does the same thing, so any Linux that uses bash as /bin/sh will also. You must be testing on a Linux that uses something else as /bin/sh. Better to use the printf command if you are worried about compatibility. echo [-e | -n] [string ...] Print a space-separated list of the arguments to the standard output and append a newline character. -n Suppress the output of the trailing newline. -e Process C-style backslash escape sequences. The echo command understands the following character escapes: -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Anyone using freebsd ZFS for large storage servers?
In the last episode (Jun 01), Wojciech Puchar said: and unbelievable narrow cases, when you don't have or can't access a backup (which you should have even when using ZFS), and you _need_ to do some forensic analysis on disks, ZFS seems to be a worse solution than UFS. On ZFS, you never can predict where the data will go. Add several disks to true. in UFS for example inodes are at known place, and flat structure instead of tree is used. even if some sectors are overwritten with garbage then fsck can scan over inodes and recover all that can be recovered. ZFS is somehow in that part similar to Amiga Fast File System. when you overwrite a directory block (by hardware fault for example), everything below that directory will disappear. You may not be even aware of it until you need that data On the other hand, even on a single-disk pool, ZFS stores two copies of all metadata, so the chances of actually losing a directory block are extremely remote. On mirrored or RAIDZ pools, you have at least four copies of all metadata. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Performance of Geli compared to Luks
In the last episode (Mar 25), Moritz Schlarb said: My HP Proliant Microserver, which I want to use as a NAS, is using an AMD Turion II Neo N40L processor. The plain disks give me about 100 MB/s using filebench with the fivestreamwrite/multistreamwrite workloads (as software RAID-1 under Linux and zpool mirror under FreeBSD). In Linux, using LUKS with cipher aes-xts-plain64 on a software RAID-1 I get the same ~100 MB/s for the same benchmarks. In FreeBSD, using a mirror zfs pool with underlying GELI with AES-XTS I get only ~50 MB/s. As the encryption algorithms should be the same, I'm wondering why FreeBSD is that slow for nearly exactly the same use case. Any ideas how I could tweak my settings in FreeBSD? As a first step, I'd try reading/writing from the raw GELI and LUKS devices to rule out performance differences due to the filesystems you're using. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: NFS locking and linux NFS server
In the last episode (Mar 25), Christoph Egger said: Hi all! I have a Linux Host (2.6.32 kernel, Debian stable) providong NFS shares. Locking files on that share works fine for linux clients [0] while it fails on a freebsd 9.0-STABLE system. The interwebs indicate there have been problems witha buggy linux implementation back in 2006 but no more hits for that problem in recent times so I assume it's fixed? root@freebsd /mnt/ 11:27 0 # kldstat -v | grep nfs 341 nfscommon 386 nfslockd 344 nfsd 385 nfssvc 342 nfs 343 nfscl 384 nfslock root@freebsd /mnt/ 11:28 0 # flock test -c ls flock: test: Operation not supported root@freebsd /mnt/ 11:31 0 # mount | grep nfs 10.70.255.8:/home/ on /mnt/ (nfs) Are you running statd and lockd (in rc.conf, rpc_statd_enable=YES and rpc_lockd_enable=YES)? Make sure that rpcinfo localhost and rpcinfo otherhost both show nlockmgr and status services. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
USA Anonymous CVS
From http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/anoncvs.html USA: anon...@anoncvs1.freebsd.org:/home/ncvs (For ssh, use ssh version 2 and no password is required.) SSH2 HostKey: 2048 53:1f:15:a3:72:5c:43:f6:44:0e:6a:e9:bb:f8:01:62 /etc/ssh/ssh_host_dsa_key.pub Example A-2. Using SSH to Check Out the src/ Tree: % cvs -d anon...@anoncvs1.freebsd.org:/home/ncvs co src The authenticity of host 'anoncvs1.freebsd.org (216.87.78.137)' can't be established. DSA key fingerprint is 53:1f:15:a3:72:5c:43:f6:44:0e:6a:e9:bb:f8:01:62. Are you sure you want to continue connecting (yes/no)? yes Warning: Permanently added 'anoncvs1.freebsd.org' (DSA) to the list of known hosts. However, when I acutally issue the command, I get a different DSA key, different IP, and it will not accept any password: # cvs -d anon...@anoncvs1.freebsd.org:/home/ncvs co src The authenticity of host 'anoncvs1.freebsd.org (96.47.72.116)' can't be established. DSA key fingerprint is 4e:bc:48:a0:e1:27:0a:62:c8:da:45:31:d4:ad:b2:00. Are you sure you want to continue connecting (yes/no)? yes Warning: Permanently added 'anoncvs1.freebsd.org' (DSA) to the list of known hosts. Password: Password: Password: Permission denied (publickey,keyboard-interactive). cvs [checkout aborted]: end of file from server (consult above messages if any) Is the USA anonymous CVS server no longer operational? Thanks, Dan ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Mapping back gptid - device
In the last episode (Feb 26), kpn...@pobox.com said: If I have this: [root@gunsight1 /etc]# ls -la /dev/gptid/ total 1 dr-xr-xr-x 2 root wheel 512 Feb 1 12:44 . dr-xr-xr-x 9 root wheel 512 Feb 1 12:44 .. crw-r- 1 root operator0, 118 Feb 1 12:44 7050cab7-4add-11e1-8919-d4bed9aca1e9 crw-r- 1 root operator0, 122 Feb 1 12:44 d78fd637-4a40-11e1-ab9b-d4bed9aca1e9 crw-r- 1 root operator0, 106 Feb 1 12:44 e1ec4071-4cfb-11e1-b4ca-d4bed9aca1e9 crw-r- 1 root operator0, 112 Feb 1 12:44 f96b3a2c-4cfb-11e1-b4ca-d4bed9aca1e9 [root@gunsight1 /etc]# How do I determine which of those /gptid/ entries corresponds to which listed in kernel messages printed at boot. I have tried gpart show -r, but it doesn't give the gptid of the entire drive -- only the partitions: = 34 877920189 mfid0 GPT (419G) 34256 1 83bd6b9d-7f41-11dc-be0b-001560b84f0f (128K) 290 44041950 2 516e7cb5-6ecf-11d6-8ff8-00022d09712b (21G) 44042240 817889280 3 516e7cba-6ecf-11d6-8ff8-00022d09712b (390G) 861931520 15988703 - free - (7.6G) Try gpart list instead, and look at the rawuuid field. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: (c)dialog working differently during boot in script called by rc system than after boot?
On Feb 23, 2012, at 11:21 AM, Robarrght wrote: (Background) I have a script that uses dialog or cdialog (system, not devel,) to prompt for network settings on my FreeBSD 8.2-RELEASE box. When I make certain changes with the script, a reboot is required. But, this would not be necessary if I ran the script during boot, before networking is setup. I've hooked up my script to the rc boot system. (Issue) The trouble is that when the rc.d system runs the script, instead of the proper dialog boxes working as expected, all I get is a cleared out black screen with a white arrow locked dead center that won't move. After boot, the script works normally. I even have the script running in getty on a virtual console, where it works like a champ. Looks like your TERM variable is not set when booting. Not a big surprise really. Try setting TERM=cons25 in your script before it calls dialog Dan So, what is the hangup? Are there dependencies that (c)dialog needs which aren't started yet? (Troubleshooting) If I simply run dialog --msgbox Test message 6 25 in the rc.d file, instead of having it call my script, I get the same bad behavior. That suggests that it is not my script, but the environment ... right? I have played with the order of when the test dialog message (my rc.d file) is ordered and even when it goes way at the end, after cron starts up, (which is getting pretty close to the end,) no improvement. I've seen this work on an older FreeBSD 6.2-STABLE box. But, when I cranked one of those up and tried the same simple dialog test directly out of the rc.d script, it yielded a black and white ascii only version of the message box. That was an improvement over what I'm getting on my current FreeBSD 8.2-RELEASE version. But, I wish I knew what the former magic was. Does anyone out there see what I'm missing? Thank you, Robarrght ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: How much space do I need on / for a 7.4 to 8 stable upgrade?
In the last episode (Feb 22), Joe Moore said: I need to upgrade a server from 7.4 stable to 8.x stable. I running buildworld as I write this, and plan to build a GENERIC kernel. The root disk partition is 248 MB which was probably the auto default size when the server was originally built. I remember having to do some scrambling during the last update to free up enough space for a new kernel and I really don't want that to happen again. I have 65MB of free space on /. Is that going to be enough? I've already moved tftpboot to /usr, cleaned out /root, /boot/kernel.old, and /tmp. I did a 5.5 - 8.1 upgrade (no intermediate installs!) on a system with a 256MB root a few years ago and didn't have any problems. As Adam said, get rid of any /boot/*/*.symbol files. With symbols, a kernel directory could be 50-70 MB, but without, you're looking at only 5-15 MB. My root is only 90MB used, so there should be quite a bit of space you should be able to free up still. Try deleting web browser cache dirs or ccache trees in ~root. What else could I clean out if I need more space? I'm thinking some executables in /rescue. ls -l shows most of them being 4MB each but that can't be right. All the files in /rescue is hardlinked to each other, so they only consume 4MB total. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Asymmetric NFS Performance
In the last episode (Feb 02), Tim Daneliuk said: Server:FBSD 8.2-STABLE / MTU set to 15000 Client:Linux Mint 12 / MTU set to 8192 NFS Mount Options: rw,soft,intr Problem: Throughput copying from Server to Client is about 2x that when copying a file from client to server. The client does have a SSD whereas the server has conventional SATA drives but ... This problem is evident with either 100- or 1000- speed ethernet so I don't think it is a drive thing since you'd expect to saturate 100-BASE with either type of drive. Things I've Tried So Far: - Increasing the MTUs - This helped speed things up, but the up/down ratio stayed about the same. - Fiddling with rsize and wsize on the client - No real difference If iostat -zx 1 on the server shows the disks at 100% busy, you're probably getting hit by the fact that NFS has to commit writes to stable storage before acking the client, so writes over NFS can be many times slower than local write speed. Setting the vfs.nfsrv.async sysctl to 1 will speed things up, but if the server reboots while a client is writing, you will probably end up with missing data even though the client thought everything was written. If you are serving ZFS filesystems, stick an SSD in the server and point the ZFS intent log at it: zpool add mypool log da3. 8GB of ZIL is more than enough, but it needs to be fast, so no sticking a $10 thumb drive in and expecting any improvement :) -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: setuid directories - or other option?
In the last episode (Feb 04), Modulok said: I have a media project directory shared with windows users via samba. Every authenticated samba user that accesses the directory is forced to the same FreeBSD user, 'foo', regardless. The group also has write-access: drwxrwxr-x 47 foofoo 2.5K Feb 4 05:42 foo/ Local shell users, however, are a problem. Ideally, I want a simliar behavior for them too i.e. Any files they create in the directory are also owned by the user 'foo'. How do I do that? (See below about setuid.) I wouldn't even care who owns the files, so long as file permission bits in this directory defaulted to 664 so every member of the group 'foo' could edit them. Can I do this without changing every user's default umask? (I want to avoid that.) Is there some kind of 'umask for this directory is blah' feature? I looked at setuid bit on directories. Sounds perfect! BUT I'll be moving to ZFS soon and from what I gather, it won't work there. I guess I could have a cron job run every minute and change offending permission bits, but that feels hacky. I think you mean the setgid bit (so that all files in the subdirectory will have group=foo), and that should work on ZFS as well. Another option might be to use ACLs to grant access to the foo group outside of the standard unix mode system: setfacl -m group:foo:rwx:df:allow /path That will grant the foo group read/write/execute access on all files under /path , regardless of the regular owner/group/umask settings. Also, make sure that the zfs aclmode and aclinherit properities on the filessytem are set to something other than discard. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: setuid directories - or other option?
In the last episode (Feb 04), Modulok said: I have a media project directory shared with windows users via samba. Every authenticated samba user that accesses the directory is forced to the same FreeBSD user, 'foo', regardless. The group also has write-access: drwxrwxr-x 47 foofoo 2.5K Feb 4 05:42 foo/ Local shell users, however, are a problem. Ideally, I want a simliar behavior for them too i.e. Any files they create in the directory are also owned by the user 'foo'. How do I do that? (See below about setuid.) I wouldn't even care who owns the files, so long as file permission bits in this directory defaulted to 664 so every member of the group 'foo' could edit them. Can I do this without changing every user's default umask? (I want to avoid that.) Is there some kind of 'umask for this directory is blah' feature? I looked at setuid bit on directories. Sounds perfect! BUT I'll be moving to ZFS soon and from what I gather, it won't work there. I guess I could have a cron job run every minute and change offending permission bits, but that feels hacky. I think you mean the setgid bit (so that all files in the subdirectory will have group=foo), and that should work on ZFS as well. Another option might be to use ACLs to grant access to the foo group outside of the standard unix mode system: setfacl -m group:foo:rwx:df:allow /path That will grant the foo group read/write/execute access on all files under /path , regardless of the regular owner/group/umask settings. Also, make sure that the zfs aclmode and aclinherit properities on the filessytem are set to something other than discard. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Identifying at jobs
In the last episode (Feb 02), Toomas Aas said: I've been curious about this for a long time. Say I have three jobs scheduled with at, as seen with atq: # atq DateOwner Queue Job# Thu Jan 12 22:12:00 EET 2012rootc 6931 Fri Jan 13 03:44:00 EET 2012rootc 6932 Fri Jan 13 04:01:00 EET 2012rootc 6933 How do I tell which job does what? I can see the files corresponding to jobs in /var/at/jobs: You want at -c job. man at: -c Cat the jobs listed on the command line to standard output. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
release 9.0 and fvwm
I recently installed fvwm (i.e. /usr/ports/x11-wm/fvwm) under FreeBSD release 9.0 and noticed some problems. Fvwm consumes all available cpu time. The GoodStuff and FvwmWinList modules don't work. Otherwise it seems to work ok. Has anyone noticed similar behavior? (Is there a fix?) The fvwm port does not seem to have changed since FreeBSD release 8.1 when it was still working correctly. Dan Strick mla_str...@att.net ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Network throughput
In the last episode (Jan 11), Ott Köstner said: On 11.01.2012 16:57, Matthew Seaman wrote: Yes -- mrtg is nice, but it relies on the snmp interface counters, and in this case with Gb traffic levels a 32bit counter will wrap in a few minutes. mrtg samples the interface counter every 5 minutes IIRC, so would probably be confused by that wrap-around happening at around the same frequency. Unless you switch to using 64bit interface counters, but I don't know if mrtg can cope with integer types that wide... Cheers, Matthew Yes, I have seen that wrapping with few hundred megabit port traffic. New snmp port has 64 bit counter option. Personally, I have already built a new snmp with 64 bit counter option, but not yet tested if it works. Maybe somebody here can comment? Definitely works. You will have to tell mrtg to make SNMPv2 requests for the 64-bit OIDs to appear, but that's it. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: little things
In the last episode (Jan 11), Gmail.com said: hello all, i like to know when freebsd foundation will be make ipfw dscp and ipfw add 1 allow all from any to any ontime 17:00 to 18:00, and help us to obtain traffic interface more easy how to linux simple ifconfig command, show us tx/rx traffic, this help us in the construction manager systems for controlling internet providers. As far as network traffic info, ifconfig is used for configuring interfaces, not monitoring them. The netstat -ib command will give you total bytes sent/received per interface, but the best solution is probably to poll them with snmp. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Exact timestamp for sorting and renaming files according to creation order
In the last episode (Jan 03), Polytropon said: For a sorting script, I'm currently searching for a method to get file creation date and time as exactly as possible. The best resolution I could get was seconds. In case more than one file is created within the same second, it doesn't work precisely enough. It should work from sh script. For the purpose of preparing the sort list (that will be sorted and then be used as a template for renaming the files with a prefix and a counter), I'm using the stat program which creates output like this: % stat -f %N %B -t %Y-%m-%d_%H:%M:%S 1.txt 1.txt 2012-01-03_12:12:12 It's also possible to use the Epoch time format, but it doesn't provide a solution better than seconds: % stat -f %N %B -t %s 1.txt 1.txt 1325589132 If you ask for the date to be printed in float (F) format, it gives more precision. The default is unsigned int (U) format. % stat -f %N %FB /COPYRIGHT /COPYRIGHT 1306190895.046721049 I've read the manuals for stat as well as for strftime (which is the facility stat's -t parameter addresses), but found nothing that is more precise than seconds. Does anyone have a suggestion how to precisely determine the order files have been created? -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Exact timestamp for sorting and renaming files according to creation order
In the last episode (Jan 03), Polytropon said: On Tue, 3 Jan 2012 14:49:02 -0600, Dan Nelson wrote: If you ask for the date to be printed in float (F) format, it gives more precision. The default is unsigned int (U) format. % stat -f %N %FB /COPYRIGHT /COPYRIGHT 1306190895.046721049 Strangely, I only get a 0 suffix for any time stamp, no matter if I create the file or apply the command as shown above to an existing file: % stat -f %N %FB /COPYRIGHT /COPYRIGHT 1313951230.0 Am I missing some file system feature? Otherwise, this _exactly_ looks like what I'm searching for. It doesn't need to be a human-readable date representation. by the way, I'm running FreeBSD 8.2-STABLE/x86 of late August 2011 here, file system used is UFS2. Try raising the vfs.timestamp_precision sysctl above zero. That gives me useful fractional values on ufs. % sysctl -d vfs.timestamp_precision vfs.timestamp_precision: File timestamp precision ( 0: seconds, 1: sec + ns accurate to 1/HZ, 2: sec + ns truncated to ms, 3+: sec + ns (max. precision)) -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: why newline scape sequence does not work in Freebsd's bash
In the last episode (Dec 31), Erik Trulsson said: On Fri, Dec 30, 2011 at 06:47:45PM -0800, Edward Martinez wrote: On 12/30/11 17:06, ??? wrote: I used ' singe quotes, so double quotes is: $ FRUIT_BASKET=apples oranges pears $ echo -e My fruit basket contains: \n $FRUIT_BASKET My fruit basket contains: apples oranges pears Thanks for the help, it worked. I find it interesting that FreeBSD's echo man page does not mention the -e option is needed to enable backslash escapes. I remembered why it worked on linux is because i created an echo alias with the -e option. So i will do the same for FreeBSD. The echo(1) manpage on FreeBSD doesn't say anything about '-e' because that version of echo doesn't have such an option. The echo you were actually using is the one builtin into bash and is described in the bash(1) manpage (including mention of the -e option.) If you want a command guaranteed to process backslash-escape sequences, use the printf command, not echo. http://pubs.opengroup.org/onlinepubs/9699919799/utilities/echo.html It is not possible to use echo portably across all POSIX systems unless both -n (as the first argument) and escape sequences are omitted. The printf utility can be used portably to emulate any of the traditional behaviors of the echo utility as follows (assuming that IFS has its standard value or is unset). New applications are encouraged to use printf instead of echo. http://pubs.opengroup.org/onlinepubs/9699919799/utilities/printf.html The %b conversion specification is not part of the ISO C standard; it has been added here as a portable way to process backslash-escapes expanded in string operands as provided by the echo utility. See also the APPLICATION USAGE section of echo for ways to use printf as a replacement for all of the traditional versions of the echo utility. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: What's wrong with this code?
In the last episode (Dec 13), John Levine said:
This tiny routine is in a .so loadable module I use. (It's part of the
mailfront SMTP daemon.)
static const char* date_string(void)
{
static char datebuf[64];
time_t now = time(0);
struct tm* tm = gmtime(now);
strftime(datebuf, sizeof datebuf - 1, %d %b %Y %H:%M:%S -, tm);
return datebuf;
}
I was getting bogus dates. Running it under GDB, time() is returning 1,
-and setting errno to 22, which is EINVAL. Changing the call to
time to time(NULL) or time(now) made no difference.
The manpage says that time() can fail for any of the reasons described in
gettimeofday(2), but time() actually calls clock_gettime(CLOCK_SECOND),
which technically could return EINVAL if the first argument isn't a valid
clock_id. CLOCK_SECOND is valid, though, so in practice it should never
fail with EINVAL. You could try adding a printf to
sys/kern/kern_time.c:kern_clock_gettime() to see if it's really failing
there.
I changed it to a call to gettimeofday(), which works fine. But what
could the problem have been? When I splice this routine into a tiny test
program that calls it and prints out the result, it works fine.
The obvious problem, since it's in a .so, is that it's linking to
something other than the system library time() function, but I did an nm
on the .so, and it said this, which sure looks like the system time()
function to me:
U time@@FBSD_1.0
Setting a breakpoint in gdb gets a complaint about trying to set a
breakpoint in /lib/libc.so.7.
Setting a breakpoint in a llibc should work fine, since time() is a regular
function and not a syscall stub. Have you built a libc with debugging
symbols? ( easy way: add DEBUG_FLAGS=-g to the top of
/usr/src/lib/libc/Makefile, and run make obj make clean make depend
make make install in that directory )
Any ideas what the problem was?
--
Dan Nelson
dnel...@allantgroup.com
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: restore(8) to UFS on USB key: terrible slow
In the last episode (Dec 08), Matthias Apitz said: I encounter the following problem with UFS file systems on USB keys, i.e. the problem is not only with one key, but with all I have; the key in question here is: Dec 7 22:17:47 tinyCurrent kernel: umass0: Generic Mass Storage, class 0/0, rev 2.00/1.02, addr 3 on usbus4 Dec 7 22:17:47 tinyCurrent kernel: umass0: SCSI over Bulk-Only; quirks = 0x4101 Dec 7 22:17:47 tinyCurrent root: Unknown USB device: vendor 0x058f product 0x6387 bus uhub4 Dec 7 22:17:48 tinyCurrent kernel: umass0:0:0:-1: Attached to scbus0 Dec 7 22:17:48 tinyCurrent kernel: da0 at umass-sim0 bus 0 scbus0 target 0 lun 0 Dec 7 22:17:48 tinyCurrent kernel: da0: Generic Flash Disk 8.07 Removable Direct Access SCSI-2 device Dec 7 22:17:48 tinyCurrent kernel: da0: 40.000MB/s transfers Dec 7 22:17:48 tinyCurrent kernel: da0: 7650MB (15667200 512 byte sectors: 255H 63S/T 975C) [...] after a while it get stuck and only does create the dirs with a frequence of one per second; also the used space is growing very slow: # df -kh /mnt Filesystem SizeUsed Avail Capacity Mounted on /dev/da0s1a7.2G 42M6.6G 1%/mnt this is after an hour or so: only 42 MBytes are written now; Any idea, what is broken with this? Cheap USB thumb drives aren't really optimized for small random-I/O writes. Can you try mounting the filesystem async? that might help a little. A workaround would be to use mdconfig to create a block device (backed by either swap or a file on your hard drive) the same size as your flash drive, newfs and restore to that, then umount the filesystem and dd the raw image directly to your flash drive. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Command which does not work anymore?
In the last episode (Dec 05), Patrick Lamaiziere said: Hello, I'm using 9.0 and the command below does not work, as far I remember that worked fine with 8.2: The goal is to copy an usb image from an OpenBSD host to an usb key on my desktop via ssh. On openbsd: # cat ucop2.img | ssh patrick@xx dd of=/dev/da0 bs=10M dd: /dev/da0: Invalid argument 0+5 records in 0+4 records out 81920 bytes transferred in 0.107747 secs (760299 bytes/sec) dd with a bs= option tells dd to use read() syscalls with a 10mb size, but ssh is going to feed it data in much smaller chunks, and they're unlikely to be multiples of a disk block in size (which you have to use if you are writing directly to a raw disk device). The 0+5 records in/0+4 records out lines say that dd read 5 blocks of data (but they were less than 10mb), and wrote 4 blocks (again less than 10mb). Try using a smaller blocksize (8k or 4k), or use a buffering program like ports/misc/team or misc/buffer just in front of your dd command, so that dd always sees block-sized writes from its stdin stream. If that command worked in 8.2, it was just by accident. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Virtualization manager suggestions
On Wed, Nov 30, 2011 at 11:05:53AM -0600, Ryan Coleman wrote: I appreciate the thought, but we have to have them on our local network. Two new projects that I've been hearing about are ganeti for (persistent) vm hosting and openstack for (transient) cloud type hosting. I do not know if freebsd will run on either of them, but hope springs eternal. dan ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: How to persist tape configuration
In the last episode (Nov 10), Andrea Venturoli said: I need to set my tape with: mt comp off. After a reboot I need to do this again. I could write a script and place it in /usr/local/etc/rc.d, but I'm just wondering if there's already something standard in /etc. A quick grep doesn't come up with anything. An rc.d script is probably your best bet. You might be able to do it with a devd.conf entry, too. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: idletime in login.conf
In the last episode (Oct 31), Damien Fleuriot said: On 30 Oct 2011, at 16:45, Modulok modu...@gmail.com wrote: List, Goal: Automatically logoff users that are idle after 'x' minutes. Attempt: I added this to /etc/login.conf to the default login class: :idletime=10m: I then rebuilt the database: cap_mkdb /etc/login.conf Problem: It doesn't work. The 'w' command shows users idle for 20 minutes or more. I'm missing something. Other than an autologout shell variables, how do I force idle users to logout? Are these users that logged in after or before your change ? Obviously the idle timeout only applies to users who logged in after you adjusted the value and rebuilt the db. Actually, the idle timeout is not implemented in the base system. It's there in case someone writes a 3rd-party module that does idle detection. See the login.conf manpage: RESERVED CAPABILITIES The following capabilities are reserved for the purposes indicated and may be supported by third-party software. They are not implemented in the base system. [...] idletime timeMaximum idle time before logout. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: came error from nc
In the last episode (Oct 31), Camal said: Hi. My name is Jamal. I am from Azerbaijan. When I tried open local port with nc and send to him /usr/local/bin/bash came error. nc -l 12345 -e /bin/bash , Error- nc: getaddrinfo: servname not supported for ai_socktype in /etc/resolv.conf and /etc/hosts files checked all of was right. Why is it so? I can't reproduce your error message, but your command won't do what you want anyway. The nc command that comes with FreeBSD doesn't support the listen on a socket and run a command option. Its -e is an ipsec option. If you install /usr/ports/net/netcat and use that command, it should work: netcat -l -p 12345 -e /usr/local/bin/bash -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: top: how to obtain which process take CPU time
In the last episode (Oct 01), ??? ??? said: hi, Freebsd-questions. last pid: 92665; load averages: 2.40, 2.68, 4.75up 5+02:45:23 20:29:07 218 processes: 3 running, 215 sleeping CPU: 59.6% user, 0.0% nice, 40.4% system, 0.0% interrupt, 0.0% idle Mem: 225M Active, 81M Inact, 115M Wired, 3936K Cache, 60M Buf, 64M Free Swap: 2048M Total, 190M Used, 1857M Free, 9% Inuse PID USERNAMETHR PRI NICE SIZERES STATETIME WCPU COMMAND 92520 cacti 1 -80 22796K 12656K piperd 0:00 1.46% php 92593 cacti 1 -80 4620K 2316K piperd 0:00 1.46% perl5.8.8 92594 cacti 1 80 3460K 1120K wait 0:00 1.46% sh 92592 cacti 1 80 3460K 1120K wait 0:00 1.46% sh 92595 cacti 1 550 5448K 2692K select 0:00 1.37% snmpget top -SIHP last pid: 99336; load averages: 1.47, 2.05, 3.66up 5+02:52:06 20:35:50 291 processes: 6 running, 266 sleeping, 18 waiting, 1 lock CPU: 52.2% user, 0.0% nice, 27.9% system, 0.0% interrupt, 19.9% idle Mem: 236M Active, 81M Inact, 115M Wired, 3436K Cache, 60M Buf, 54M Free Swap: 2048M Total, 189M Used, 1858M Free, 9% Inuse PID USERNAME PRI NICE SIZERES STATETIME WCPU COMMAND 11 root 171 ki31 0K 8K RUN 70.1H 27.59% idle: cpu0 98502 cacti80 23820K 12932K nanslp 0:00 0.20% php 44054 root 80 3124K 524K nanslp 0:56 0.10% monitord 99051 root440 3496K 2020K RUN 0:00 0.10% top 99331 cacti -80 4620K 2332K piperd 0:00 0.00% perl5.8.8 99326 cacti -80 4620K 2332K piperd 0:00 0.00% perl5.8.8 It is unclear which process take CPU time. is there any other tool, which help me to see processes that take CPU? Take a look at the last pid values and the time of each top snapshot: last pid: 92665; load averages: 2.40, 2.68, 4.75up 5+02:45:23 20:29:07 last pid: 99336; load averages: 1.47, 2.05, 3.66up 5+02:52:06 20:35:50 So, in 7 minutes your system has launched 6671 processes, and in your first top snapshot I see sh and snmpget processses. The reason you don't see any processes in top consuming CPU is that they are being created and then exit before top can print them. If you turn process accounting on (add accounting_enable=YES to /etc/rc.conf, then run /etc/rc.d.d/accounting start), you can run the lastcomm command to see a log of every process that has exited and its CPU usage. It won't really tell you any more than you are running Cacti and monitoring a lot of devices :) -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Overwrite bsd tar
In the last episode (Sep 23), Esimorp E said: Hi Bernt, Thank you for your reply. I had similar concern as well but I really needed the globus toolkit and wouldn't mind if its the only that works on the system. You never actaully said what the problem was. If it's a bug in bsdtar, I'm sure it could get fixed quickly given a sample tarfile. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Segmentation fault, _malloc_prefork () - debugging help needed
In the last episode (Sep 18), Unga said: I'm developing an multi-threaded application on FreeBSD. When it is running for sometime, it develops a Segmentation fault. The ddd debugger shows following: Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 296c6580 (LWP 100137)] 0x28ee390e in _malloc_prefork () from /lib/libc.so.7 How could I know the exact line in source where this issue develops? If you have a full /usr/src tree extracted, you can edit /usr/src/lib/libc/Makefile and add DEBUG_FLAGS=-g at the top, then run make obj make depend make make install to install the new libc with debugging symbols. Then your debugger will show more info for functions inside libc. If you don't have a source tree checked out yet, install the devel/subversion-freebsd port, cd into /usr/src/ and run svn co svn://svn.freebsd.org/base/stable/8 . (or base/release/8.2.0, or base/stable/7 or base/head, depending on which version you want; you can browse the branches at http://svnweb.freebsd.org/base/ before you checkout anything) -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Accounting disabled/enabled messages
In the last episode (Sep 05), David J. Weller-Fahy said: I noticed today the following entries in my dmesg. #v+ Accounting enabled Accounting disabled Accounting enabled Accounting disabled Accounting enabled Accounting disabled Accounting enabled #v- The uname -a follows. FreeBSD NAStie 9.0-BETA1 FreeBSD 9.0-BETA1 #0: Thu Jul 28 16:34:16 UTC 2011 r...@obrian.cse.buffalo.edu:/usr/obj/usr/src/sys/GENERIC i386 I tracked the actual messages down to /usr/src/sys/kern/kern_acct.c, but am not familiar enough with kernel internals to figure out why its happening. It *looks* like its happening every time I reboot the system (been playing around with the power). However, I want to be sure, and figured I'd check with the list. So - are these messages something to be concerned about? Or something to ignore? Or something to ignore in a BETA environment, and not in a production environment? Assuming you have accounting enabled in /etc/rc.conf, that's to be expected. Accounting is enabled on boot, disabled on shutdown, and cycled twice during /var/account/acct rotation at 3am. See /etc/rc.d/accounting and /etc/periodic/daily/310.accounting . -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: nss-3.12.11.with.ckbi.1.87.tar.gz: File unavailable (e.g., file not found, no access)
In the last episode (Sep 05), Antonio Olivares said: Dear folks, sorry to bother you guys, but I am encountering a problem updating I need 4 ports only, but can't get past the error above: Building new INDEX files... done. === New version available: ca_root_nss-3.12.11_1 === New version available: gtk-2.24.6 === New version available: gtk-update-icon-cache-2.24.6 === New version available: firefox-6.0.1,1 === 402 total installed ports === 4 have new versions available grullahighschool# portmaster -a === License check disabled, port has not defined LICENSE === Found saved configuration for ca_root_nss-3.12.9 = nss-3.12.11.with.ckbi.1.87.tar.gz doesn't seem to exist in /usr/ports/distfiles//. = Attempting to fetch http://ftp.mozilla.org/pub/mozilla.org/security/nss/releases/NSS_3_12_11_WITH_CKBI_1_87_RTM/src/nss-3.12.11.with.ckbi.1.87.tar.gz fetch: http://ftp.mozilla.org/pub/mozilla.org/security/nss/releases/NSS_3_12_11_WITH_CKBI_1_87_RTM/src/nss-3.12.11.with.ckbi.1.87.tar.gz: No address record This is your main problem; you aren't able to resolve ftp.mozilla.org for some reason. The other sites are mirrors that either aren't mirroring the security subdirectory, or haven't updated their mirror recently enough to have a copy of that file. I tried to get the file manually, but it does not exist. Thanks for advice/suggestions/comments. It definitely does exist at the above url. Since you seem to be having DNS issues, try putting 63.245.209.137 ftp.mozilla.org in your /etc/hosts file and try fetching again, since that's what ftp.mozilla.org currently resolves to. Remember to remove the line after fetching, since the IP may change later. http://www.robtex.com/dns/ftp.mozilla.org#records -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: wpa_cli issues
In the last episode (Aug 19), Zane C. B-H. said: On Fri, 19 Aug 2011 09:14:54 -0500 (CDT) Robert Bonomi bon...@mail.r-bonomi.com wrote: From: Zane C. B-H. v.ve...@vvelox.net Is there any way to undefine a variable once it has been set? *As(stated*, the answer involves the offspring of the mating of a rhinoceros and an elephand. =GUESSING= that you mean a shell 'envionment variable', the answer is 'yes'. _How_ one can do it depends on the shell (*unspecified*!) being used. 'unsetenv' _may_ do the trick. Alternatively a variable assignment with no value (.e.g VARIABLE=) may work. Blarg? None of these is even vaguely related to my question about wpa_cli, as stated in the subject. wpa_cli only understands a fixed list of variables to set, and it doesn't make sense to undefine them. You can set them back to their default values, but they must have a value. Defaults from looking at the source: EAPOL::heldPeriod = 60 EAPOL::startPeriod = 30 EAPOL::maxStart = 3 EAPOL::authPeriod = 30 dot11RSNAConfigPMKLifetime = 43200 dot11RSNAConfigPMKReauthThreshold = 70 dot11RSNAConfigSATimeout = 60 Running set from within wpa_cli should print these values, too, according to the manpage. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: dtrace function arguments
In the last episode (Aug 15), Ashley Williams said: I'm looking for a faster way to get more verbose information about dtrace function arguments. For example. Say, I want to know more about the funciton syscall:freebsd32:connect:return. I'd start off by doing a listing: # dtrace -lvf connect [...] Argument Types args[0]: int args[1]: caddr_t args[2]: int From the output of the listing, I can see quite clearly there are three arguments for this function - int, caddr_t, int; but I can't see from this output what these refer to. I could probably find the answer by digging through header files and source code, but this isn't exactly efficient. Is there an easier way to find more information about functions (not specifically this one)? All syscalls should have a manpage documenting their arguments, and some common kernel functions have manpages in section 9 (so man 9 malloc will get the kernel version, for example), but most kernel functions aren't officially documented apart from comments in the source. http://fxr.watson.org/ is a handy resource for finding where in the source tree a given function is defined. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Group permissions are broken?
In the last episode (Aug 15), Yuri said: User john is a member of both webcamd and vboxusers: # grep john /etc/group webcamd:*:145:john vboxusers:*:920:john When the file /tmp/my-test is owned by webcamd, user john can touch it ok: $ ls -l /tmp/my-test ; touch /tmp/my-test -rw-rw 1 vboxusers vboxusers 0 Aug 15 12:54 /tmp/my-test But when /tmp/my-test is owned by webcamd, user john gets an error: $ ls -l /tmp/my-test ; touch /tmp/my-test -rw-rw 1 webcamd webcamd 0 Aug 15 13:02 /tmp/my-test touch: /tmp/my-test: Permission denied Why does this error occur? Two groups seem identical. Just different group ids. Filesystem is UFS: /dev/ad10s1a on / (ufs, NFS exported, local) As a sanity check, what is the output of the groups command? If you recently edited /etc/group, maybe you need to log out and back in to a credential with the new group list? -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: listing ZFS pool name in fixit mode
In the last episode (Aug 05), claudiu vasadi said: I have a 8.2-amd64 machine on which I blew up the boot blocks (isn't booting any more). It's s supermicro server btw and we have a ZFSonRooT setup there. The way I want to go about fixing it is booting into single user mode, mounting all the datasets, undoing the modification I did to the boot blocks, recompiling and installing the old version of the boot blocks for the system to boot. The problem is that after I boot a 8.2-DVD and then go to fixit mode, load the opensolaris and the zfs modules, I cannot list the ZFS pool name (I don;t know the ZFS pool name because I didn;t setup this machine) and without the name, I cannot import the pool to fix the boot blocks. So, my question is if there is a way to list the zpool in single user mode so that I know which pool I have to import or if there might be another way of fixing this problem. zpool import with no pool name will list all the pools available to import. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: gmake[4]: *** [Gdk-2.0.gir] Error 127 = updating gtk to latest on ports
In the last episode (Aug 04), Antonio Olivares said: On Thu, Aug 4, 2011 at 8:00 PM, Warren Block wbl...@wonkity.com wrote: On Thu, 4 Aug 2011, Antonio Olivares wrote: In an effort to keep up to date, I checked updates that are available and tried to apply them. Encountered a problem with gtk : /* Commands run */ quadcore# . Looking up portsnap.FreeBSD.org mirrors... 5 mirrors found. Fetching snapshot tag from portsnap5.FreeBSD.org... done. Latest snapshot on server matches what we already have. No updates needed. Ports tree is already up to date. === New version available: gtk-2.24.5_1 === 537 total installed ports === 1 has a new version available ran # portmaster -a Always (yes, always) check /usr/ports/UPDATING before throwing any automatic update tool at it. See the 20110730 entry. Yes I see it, but it does not make a difference :( 20110730: AFFECTS: users of x11-toolkits/gtk20 AUTHOR: gn...@freebsd.org The gtk-update-icon-cache utility has been split out of the gtk20 port. Use the following instructions to update your system. # pkg_delete -f gtk-2.\* # portmaster x11-toolkits/gtk20 # portmaster -a I ran the first command successfully, but the second bombs out the same place and with same error :( ual-x11.c x11/gdkwindow-x11.c x11/gdkxftdefaults.c x11/gdkxid.c x11/xsettings-client.c x11/xsettings-common.c libgdk-x11-2.0.la Makefile --output Gdk-2.0.gir /usr/local/bin/g-ir-scanner: not found g-ir-scanner is a python script. I bet the interpreter path on the first line of that file no longer points to a valid python executable. Have you converted python versions recently and forgot to rebuild all ports depending on it? -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: larger disk for a zfs pool
In the last episode (Aug 01), Dick Hoogendijk said: OK, my freebsd system runs on ZFS boot. W/ solaris getting larger disks for a pool was quit easy. Simply replace one disk from a mirror for a larger one, wait for the resilvering and after this replace the second one for a larger disk and wait for the resilvering again. That's it. Been there, done that. But my feeling tells me it is not that simple for a FreeBSD zfs root system, or is it? Should be the same procedure. Make sure you either use zpool online -e when swapping in the new disks, or that you have the zpool autoexpand=on attribute set. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: devel/libnotify: Error: shared library notify.1 does not exist
In the last episode (Jul 26), Peter Ulrich Kruppa said: building www/firefox fails with Error: shared library notify.1 does not exist Is this a problem with firefox- or libnotify-port? And what can be done? What version of the libnotify and firefox ports are you trying to build (current are libnotify-0.5.2 and firefox-5.0,1)? Do you have a /usr/local/lib/libnotify.so.1 file? If not, try reinstalling the libnotify port. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: high performance open source DHCP solution?
In the last episode (Jul 19), Rogelio said: The free DHCP solution, ISC, seems to be having scaling issues (i.e. handling only about 200 DHCPDISCOVER and 20 DHCPRENEW requests), and I was wondering if anyone had any open source suggestions of solutions that could scale much better? 200 per second sounds like pretty good performance, but you haven't given any info about your setup either. Are you disk bound or CPU bound at this point? -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Am I Missing A Compat Library?
In the last episode (Jul 18), Tim Daneliuk said:
On 7/18/2011 12:43 PM, Roland Smith said this:
But I _don't_ get this linker error. Just a thought, but have a look at
glib and dbus, and check that it is linked to the right libicu? ('ldd
/usr/local/lib/libglib-2.0.so.0' and 'ldd
/usr/local/lib/libdbus-glib-1.so.2')
If that is not the case, you'll have to dive into the configure output,
I think.
A forced deinstallation of avahi-app and then an installation, seems to
have made the problem disappear.
So much for dynamically liked libraries will make systems maintenance
simpler ...
BTW, Both portupgrade and portmaster (with the -w flag) will copy old shared
libraries to /usr/local/lib/compat/pkg/ during an upgrade, so that programs
linked to them will still run.
--
Dan Nelson
dnel...@allantgroup.com
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: how to force a hard reboot remotely
In the last episode (Jul 16), Aryeh Friedman said: Is there any way to force a complete power down and then reset of a i386 without physically being present? If you have a server motherboard with IPMI (or a high-end server with a service processor module), you can use that to power cycle the system remotely even if the OS is hung. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: IPFW Firewall NAT inbound port-redirect
In the last episode (Jul 12), Michael Sierchio said: Is there a way of specifying a particular public address if there is more than one bound to the external interface? A la nat 123 config if re0.2 log same_ports redirect_port tcp 10.0.0.3:22 102.10.22.1: Yes; the redirect_port syntax is described in the natd manpage: redirect_port proto targetIP:targetPORT[-targetPORT] [aliasIP:]aliasPORT[-aliasPORT] [remoteIP[:remotePORT[-remotePORT]]] -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: IPFW Firewall NAT inbound port-redirect
In the last episode (Jul 12), Michael Sierchio said: On Tue, Jul 12, 2011 at 9:03 AM, Dan Nelson dnel...@allantgroup.com wrote: In the last episode (Jul 12), Michael Sierchio said: Is there a way of specifying a particular public address if there is more than one bound to the external interface? A la nat 123 config if re0.2 log same_ports redirect_port tcp 10.0.0.3:22 102.10.22.1: Yes; the redirect_port syntax is described in the natd manpage: redirect_port proto targetIP:targetPORT[-targetPORT] [aliasIP:]aliasPORT[-aliasPORT] [remoteIP[:remotePORT[-remotePORT]]] We're not talking about natd. The question was about the use of ipfirewall nat. Right, but ipfw nat is basically the userland libalias library loaded as a kernel module, so the config parameters are the same. $ grep MODULE_DEPEND /sys/netinet/ipfw/ip_fw_nat.c MODULE_DEPEND(ipfw_nat, libalias, 1, 1, 1); MODULE_DEPEND(ipfw_nat, ipfw, 2, 2, 2); also, man ipfw: NETWORK ADDRESS TRANSLATION (NAT) ipfw support in-kernel NAT using the kernel version of libalias(3). [..] REDIRECT AND LSNAT SUPPORT IN IPFW Redirect and LSNAT support follow closely the syntax used in natd(8). See Section EXAMPLES for some examples on how to do redirect and lsnat. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: DNS and file system messed up...
On Jul 8, 2011, at 9:54 PM, Gary Kline wrote: On Fri, Jul 08, 2011 at 07:27:12AM -0600, Dan Busarow wrote: Gary, add named_flags=-c /etc/namedb/named.conf to /etc/rc.conf. Or change /etc/namedb/named.conf to the /var version if you like/there is no symlink. Dan Dan! I think you fixed something. I haven't figured this out yet, and would be grateful if you could decode this in /var/log/messages:: Jul 8 20:39:32 ethic named[83003]: stopping command channel on :: 1#953 Jul 8 20:39:32 ethic named[83003]: exiting Jul 8 20:39:37 ethic named[84090]: starting BIND 9.3.6-P1 -c /etc/namedb/named.conf -t /var/named -u bind Jul 8 20:39:37 ethic named[84090]: none:0: open: /etc/rndc.key: file not found Gary, Theres probably an /etc/rc.conf line to fix these but what I always do is simply symlink /etc/namedb/rndc.key to /etc/rndc.key # ln -s /etc/namedb/rndc.key /etc/rndc.key I actually use rndc.conf on my systems but I think the names and files are interchangeable. Dan Jul 8 20:39:37 ethic named[84090]: couldn't add command channel 127.0.0.1#953: file not found Jul 8 20:39:37 ethic named[84090]: none:0: open: /etc/rndc.key: file not found Jul 8 20:39:37 ethic named[84090]: couldn't add command channel :: 1#953: file not found Jul 8 20:39:37 ethic named[84090]: the working directory is not writable Jul 8 20:39:37 ethic named[84090]: running This, after I added your named_flags line into /etc/rc.conf. Where I get lost is *what* gives me that none:0 lines?? I see the same or worse err when I drop in bind98. IIRC, named does run, but the messages log is fulll of rndc.key error messages that I just cannot understand. _Now_, having dropped in your named_flags line, I am seeing something similar. I haved grepped thru the entire /etc/ tree and haven't found anything that explains where I messed up Ideas? thanks to you or anybody else onlist. gary ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: DNS and file system messed up...
On Jul 8, 2011, at 3:01 AM, Matthew Seaman wrote: On 08/07/2011 08:25, Doug Hardie wrote: On 7 July 2011, at 22:58, Gary Kline wrote: Jul 7 10:16:33 ethic named[54366]: none:0: open: /etc/ named.conf: file not found Jul 7 10:17:56 ethic named[54371]: starting BIND 9.3.6-P1 -c / var/named/etc/namedb/named.conf The first one that fails is looking for /etc/named.conf. The second one shows its in /var/named/etc/named/named.conf Those are different locations. I suspect you have named_flags setup in rc.conf pointing to /etc/namedb/named.conf rather than the right location. Its also possible that its not set in rc.conf but defaults in either the rc script or /etc/rc.d/named. On my system it appears to default in /etc/rc.d/named. FreeBSD defaults to running named chrooted. /etc/namedb is actually a symbolic link: % ls -la /etc/namedb lrwxr-xr-x 1 root wheel 21 Jul 6 06:24 /etc/namedb@ - /var/named/etc/namedb so the files referenced are in fact exactly the same file. Actually /etc/named.conf is NOT the same as /etc/namedb/named.conf ergo it is not the same as /var/named/etc/ namedb/named.conf Gary, add named_flags=-c /etc/namedb/named.conf to /etc/rc.conf. Or change /etc/namedb/named.conf to the /var version if you like/there is no symlink. Dan However, the flags from the log extract don't look like the defaults to me. (I'm running the dns/bind98 port, and the equivalent info from the log line is '-t /var/named -u bind') Gary, what named related settings do you have in /etc/rc.conf? You almost certainly don't need anything more than: named_enable=YES and perhaps syslogd_flags=-ss -l /var/named/var/run/log so named can log to the system syslog. Cheers, Matthew -- Dr Matthew J Seaman MA, D.Phil. 7 Priory Courtyard Flat 3 PGP: http://www.infracaninophile.co.uk/pgpkey Ramsgate JID: matt...@infracaninophile.co.uk Kent, CT11 9PW ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: import users from LDAP to local password file (pwd.db)
In the last episode (Jul 08), Frank Bonnet said: On 07/08/2011 05:43 PM, Moises Castellanos wrote: On Fri, Jul 8, 2011 at 9:56 AM, Frank Bonnetf.bon...@esiee.fr wrote: I need to import the necessary users's data from an OpenLDAP directory server to put them in the local password files , anyone has done this before ? The machine use nss_ldap and pam_ldap to authenticate users but for robustness during the holidays I would like to have a local password file on this machine which is our mailhub. The OpenLDAP server runs on another machine and if it fails during holidays I want my mailhub to be standalone for authentication in order to let email service running even the directory server crash. You can try with getent(1) passwd and see if you can work with the output getent does not show the encrypted password field LDAP servers usually don't allow clients to see the raw password hash. Authentication checks are done by binding as the requested user, so the calling app doesn't have a chance to grab the hash and do an offline brute-force attack on it. To ensure LDAP availability, the usual thing to do is set up multiple servers with LDAP replication between them, and configure your client's ldap.conf to use all of them (or use carp or some other IP management app to provide a single always-up IP address). -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: PID 11 using 400% CPU
In the last episode (Jul 05), manish jain said: On 30 June 2011 10:26, Dan Nelson dnel...@allantgroup.com wrote: In the last episode (Jun 30), Manish Jain said: I have a strange problem with my 8.1 box. After booting, the hard disk goes into a full-speed never-ending spin. To see what disk I/O is being done, try running ktrace -dip 0 ; sleep 10 ; ktrace -C, to capture all syscalls done on the entire system (pid 0 plus children) for 10 seconds, then run kdump -m64 | less to view the results. Look for read or write calls. It looks like ppp is doing a lot of read and write operations, which keeps the disk spinning. How do I set this right ? Is there something wrong with my ppp.conf (see below) ? I bet that if you ran fstat or lsof on the ppp process, all the writes are actually to your serial device or a tun device, not to disk. ppp is unlikely to cause much disk I/O. You'll have to filter out the ppp process and check your kdump output again. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: PID 11 using 400% CPU
In the last episode (Jun 30), Manish Jain said: Hello All, I have a strange problem with my 8.1 box. After booting, the hard disk goes into a full-speed never-ending spin. To see what disk I/O is being done, try running ktrace -dip 0 ; sleep 10 ; ktrace -C, to capture all syscalls done on the entire system (pid 0 plus children) for 10 seconds, then run kdump -m64 | less to view the results. Look for read or write calls. 'ps waux' always shows pid 11 as taking 400% CPU utilization : /root # ps -up 11 USER PID %CPU %MEM VSZ RSS TT STAT STARTED TIME COMMAND root11 400.0 0.0 032 ?? RL7:22PM 166:35.46 [idle] I have tried multiple tweaks to resolve this - all to no effect. The As for this, what's to resolve? The idle process is a placeholder with one thread per CPU that accounts for time the CPU isn't doing any work. If you want to reduce it's CPU use, run other CPU-intensive processes :) BTW, Windows has the same thing if you look at task manager; it's called System Idle Process there. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Performance of a USB ZIL for ZFS
In the last episode (Jun 26), Joshua Isom said: On 6/25/2011 9:32 PM, Damien Fleuriot wrote: On 25 Jun 2011, at 19:17, Joshua Isomjri...@gmail.com wrote: I was wondering if anyone had tried using a decent USB flash drive for the ZIL. I know it'd be hard finding one fast enough, but some from patriot seem like they might be suitable for home use. Part of the idea is to just minimize hard drive thrashing and the wear and tear associated with it. If it helps prevent the drives from going bad, and doesn't hurt performance too bad all the better. But if it's going to hurt performance too much or not help prevent thrashing there isn't a point. I stopped reading at the title. The answer is no. Grab a SSD for $80-120ish. Perhaps it would have helped to read the email. Part of the concern is making sure the drives don't fail and not just throughput. Given that Kingston sells an SATA SSD for $40 that only gets writes at 30mb/s write, and some USB drives might get up to 20mb/s. If I get two drives and put them on different controllers, mirrored, I might get acceptable performance. I may still loose performance, but if my drives last a year longer, I can probably accept it. I'm ok with loosing some performance, but I just don't want it dragging down the system. And if it won't help the drives last longer there's no point. A seaparate ZIL isn't meant to extend the lifetime of the hard drives; it's meant to accelerate the speed of sync writes. Those are pretty infrequent themselves, unless you're an NFS server. You'll see a couple syncs per commit on a database server, but compared to the amount of regular reads and writes on your average system, you'll save under 1% of the writes by adding a fast ZIL. And remember, the ZIL is just a write log. Everything that gets written to it will get flushed to disk when zfs writes the next transaction group. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: pkg-config --cflags glib-2.0 gives wrong -I dir
In the last episode (Jun 25), Matthias Apitz said:
Why gives
$ pkg-config --cflags glib-2.0
-I/usr/local/include/glib-2.0 -I/usr/local/lib/glib-2.0/include
$ ls -ld /usr/local/include/glib-2.0 /usr/local/lib/glib-2.0/include
ls: /usr/local/lib/glib-2.0/include: No such file or directory
drwxr-xr-x 5 root wheel 512 May 28 19:01 /usr/local/include/glib-2.0
a non existing -I directory? This concrete example is with a 9-CURRENT
and glib-2.0 from the ports as glib-2.26.1_1, but I see this as well in
some older 8.x systems;
Checking Solaris and SUSE Linux, I see a similar pair of directories:
solaris$ pkg-config --cflags glib-2.0
-I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include
linux$ pkg-config --cflags glib-2.0
-I/usr/include/glib-2.0 -I/usr/lib64/glib-2.0/include
The /usr/lib* directory on each system contains a single file: glibconfig.h.
On FreeBSD, this file is in /usr/local/include/glib-2.0/ along with all the
other headers (headers don't belong in /lib/ anyway).
Try putting the attached patch into the files directory of the glib20 port.
--
Dan Nelson
dnel...@allantgroup.com
No headers are installed into ${libdir}/glib-2.0/include, so remove it from
CFLAGS
--- glib-2.26.1/glib-2.0.pc.in 2009-03-31 18:04:20.0 -0500
+++ glib-2.26.1/glib-2.0.pc.in 2011-06-25 19:14:23.580424986 -0500
@@ -12,4 +12,4 @@
Version: @VERSION@
Libs: -L${libdir} -lglib-2.0 @INTLLIBS@
Libs.private: @ICONV_LIBS@
-Cflags: -I${includedir}/glib-2.0 -I${libdir}/glib-2.0/include
@GLIB_EXTRA_CFLAGS@
+Cflags: -I${includedir}/glib-2.0 @GLIB_EXTRA_CFLAGS@
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: License for console fonts
In the last episode (May 09), Sean Hamilton said: What is the license on the fonts in /usr/share/syscons/fonts, especially swiss-8x8? Who/what is their origin? Would it be acceptable to include these binaries with a GPL-licensed program? I submitted the swiss fonts back in 1997, and I have a file called swisst.com on an old DOS directory dated 1995, so they're probably even older than that. I seem to remember the 8x16 font was inspired by a sample font that came with the ancient DOS font editor Font Mania, but for nostalgia's sake I goooled for it, started up DOSBox, and all of the sample fonts are fancy and not suitable for daily use. I don't think I came up with the name swiss myself, so I probably used another VGA font as a template, but my guess is the original author isn't concerned about the copyright status... You can consider them 2-clause BSD licensed like the rest of FreeBSD, until someone can find an older font that looks similar. I mainly wanted a sans-serif font that didn't waste so many pixels on whitespace between letters. Capitals and numbers on the 8x16 font are 1 pixel taller than a default ROM font, which really makes it stand out. The 8x8 font was copied from the 8x16 font, but at 8x8 there really isn't much personalization you can do, apart from decide serifs, or not? :) -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: ZFS Striping and Optimizing Capabilities
In the last episode (Apr 09), Chris Telting said: Just a few questions about what ZFS actually does. So if anyone has intimate knowledge about ZFS's implementation on Freebsd I'm sure I and others would appreciate the answers. When you add a second and or thrid drive/partition to a zpool I'm assuming that it's going to start using the drives like a raid 0 stripe. How do the ZFS versions differ in this? Does it immediately start striping all files in the background on low priority or does it do it as files are accessed? Does ZFS in any way do performance testing of Currently ZFS never moves a block once it has been written. The whole implementation of copy-on-write and snapshots relies on this. If you add a new zvol to a pool, new data will get balanced across all the zvols but old data will stay where it was. Ideally you should expand pools before they get too full, or zpool export/import them into new pools if you want good distribution over all disks. There have been hints for the last few years of a bp rewite feature appearing, but it has never materialized. -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Place to install library of shell functions
In the last episode (Apr 05), Michael Grünewald said: today I come to you with what seems to be somehow pedantic question: where is the best place to install libraries of shell functions. I read hier(4) carefully and it seems the correct place for this would be somewhere under `/usr/local/share': share/architecture-independent files On the base systems, many macros (make templates and groff macros) are actually installed under /usr/share. However, the only library of shell functions I am aware of is `/etc/rc.subr', that landed on this funny location, probably to be pretty sure that this library is on the same filesystem as the scripts under /etc/rc.d. Several of the ports install shell scripts under `/usr/local/lib' that hier(4) devotes to ``shared and archive ar(1)-type libraries''. These shell scripts are: The zsh port installs all its helper scripts and functions into /usr/local/share/zsh , and the portupgrade and portmaster ports insert their zsh autocompletion functions into /usr/local/share/zsh/site-functions. Seems to work great. Ports that install scripts into /usr/local/lib probably either don't have separate script and lib install paths in their makefile (tcl probably), or their scripts aren't meant to be called directly (firefox). -- Dan Nelson dnel...@allantgroup.com ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
RaLink RT2760 + RT2720
Hello, I have purchased this wireless card: http://www.monoprice.com/products/product.asp?c_id=105cp_id=10501cs_id=1050103p_id=5338seq=1format=3#specification It seems that it is not being detected by 8.2-RELEASE. I have tried the RT2870 and RT2860 drivers with the card from: http://repo.or.cz/w/ralink_drivers.git This has resulted in little success. Here is a snippet of pciconf -lvcb: none2@pci0:3:1:0: class=0x028000 card=0x30621814 chip=0x30621814 rev=0x00 hdr=0x00 vendor = 'Ralink Technology, Corp.' class = network bar [10] = type Memory, range 32, base 0xe200, size 65536, enabled cap 01[40] = powerspec 3 supports D0 D3 current D0 If anyone could be of assistance it would be greatly appreciated. My ultimate goal is to use this as an AP. Thanks, Dan ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
