Re: [gentoo-user] Good file system that recovers from a power failure.
101231 Alan McKinnon wrote: 101231 Dale wrote: What is a good file system that recovers well from a improper shutdown? Down here we have Africa power. Africa power makes post-Katrina power look tame. We get brief power outages here regularly in downtown Toronto, typically in the very early morning when their doing maintenance: no warning, no lights, no computer, back to the 1700s. Total corruptions in 5 years with reiserfs-3.6 and NO ups in that environment = zero. Same here: I've used Reiserfs since 2003 never had a problem; I installed it in my netbook 2009, as it was so stable. It seems version 3 is being lightly maintained needs no improvements; as for version 4, it's unlikely to get into the Linux kernel, as Btrfs -- or Zfs -- are more strongly favored by its devs. -- ,, SUPPORT ___//___, Philip Webb ELECTRIC /] [] [] [] [] []| Cities Centre, University of Toronto TRANSIT`-O--O---' purslowatchassdotutorontodotca
Re: [gentoo-user] Good file system that recovers from a power failure.
110101 Philip Webb wrote: typically in the very early morning when their doing maintenance: Sorry, it's early in my day: that sb they're (red face). -- ,, SUPPORT ___//___, Philip Webb ELECTRIC /] [] [] [] [] []| Cities Centre, University of Toronto TRANSIT`-O--O---' purslowatchassdotutorontodotca
Re: [gentoo-user] Re: nvidia update problems
On Friday 31 December 2010 21:00:57 Alan McKinnon wrote: Apparently, though unproven, at 20:54 on Friday 31 December 2010, walt did opine thusly: So, if you've ever run the NVIDIA install program manually, you probably have two different nvidia.ko files in /lib/modules, and the wrong one gets loaded automatically at boot time. Re-emerging nvidia-drivers will build *and* load the correct kernel module each time you do it. May not be your problem but it's easy to check, at least. I didn't think I'd ever run the nVidia installation program, and on checking I see one nvidia.ko for each kernel version I have installed, so that isn't the problem. module-rebuild takes care of all that nicely. And it wants to rebuild the module again, even though I only remerged it yesterday. Something is messing about with the nVidia module. Thanks for the ideas, gents. -- Rgds Peter. Linux Counter 5290, 1994-04-23.
Re: [gentoo-user] Core i7 M620 power management problem
Am 2010-12-31 11:59, schrieb Mick: Hmm ... could it be a buggy BIOS? Are you running the latest firmware for it? Yes, that would also have been my next question. Maybe you even *find* a bug in that BIOS right now that should be corrected.
Re: [gentoo-user] Good file system that recovers from a power failure.
On Friday 31 December 2010 22:11:37 Volker Armin Hemmann wrote: reiser4 is fully atomic. A transactions happens completely or it doesn't. Unlike ext4 or btrfs or xfs. reiser4 also uses barriers (the others use them too, but), when barriers are not available for some reason or another, it complains in dmesg and goes into sync mode. This combined makes it pretty robust against power failures. You won't get the good old xfs/ext4/btrfs problem that a rename can end with two useless empty files. I read what Volker is saying and it sounds impressive from a fs design perspective, but my experience does not concur with it. I have a had a power cut (run out of battery) and the fs got corrupted. :-( This however may not be a conclusive finding. I was running reiser4 for about a year on my laptop. Unfortunately, I have a had a large number of fs corruptions, most of which appeared to be random. The last one just before Christmas proved to be fatal and unrecoverable with fsck.reiser4. I could try to blame the disk, but the MSWindows ntfs which I dual boot to from the same disk never failed or corrupted (admittedly though it has seen hardly any use). I have to say that when it did run without corruption reiser4 is an exceptional fs in terms of performance. With the exception of mounting large partitions which takes some time, I don't think anything else I've tried comes close. If I were to build a desktop which unlike a laptop is not bounced around when commuting on trains and what not, I would probably try it again (because I have a niggling suspicion that the cause of my problems might have been a mechanical reason). Either way, I can say with some certainty that power cuts and running out of space on a partition brought about fs corruption with reiser4. I have now moved all but one of my partitions to ext4 and will wait to see what happens with that, but 11 months on reiser4 has left a bad taste in my mouth. Historically, I have mostly used reiserfs and xfs. Reiserfs is in my experience very reliable and easily recoverable and I can assuredly echo Alan's findings. Some years ago I had a faulty memory controller which would hard lock an old desktop. At least once a day (typically in the middle of an emerge, or updatedb) it would crash badly and I would have to pull he plug. In as many as 4 years I must have had hundreds and hundreds of hard reboots. I vaguely recall one or two fs corruptions on only one or two partitions. reiserfsck did recover the fs every single time without major drama. If bleeding edge performance is not an issue I would recommend reiserfs to mitigate the risk from powercuts. -- Regards, Mick signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] Good file system that recovers from a power failure.
Hi, there is one scenario that is troublesome: mldonkey temp files. If you don't use mldonkey/amule/$whatever you should be fine.
Re: [gentoo-user] Good file system that recovers from a power failure.
Philip Webb wrote: 110101 Philip Webb wrote: typically in the very early morning when their doing maintenance: Sorry, it's early in my day: that sb they're (red face). That's OK. I got a cell phone too. I do pretty good with text speak. lol Dale :-) :-)
Re: [gentoo-user] Core i7 M620 power management problem
I actually am running the latest firmware. I had thought that maybe that was the problem, but I rev'ed it about a month ago and it did not solve it. Am waiting for the Ubuntu 10.10 to finish downloading and give that a whirl. On Sat, Jan 1, 2011 at 3:16 AM, Stefan G. Weichinger li...@xunil.at wrote: Am 2010-12-31 11:59, schrieb Mick: Hmm ... could it be a buggy BIOS? Are you running the latest firmware for it? Yes, that would also have been my next question. Maybe you even *find* a bug in that BIOS right now that should be corrected. -- Bill Longman
Re: [gentoo-user] Re: nvidia update problems
Apparently, though unproven, at 12:36 on Saturday 01 January 2011, Peter Humphrey did opine thusly: On Friday 31 December 2010 21:00:57 Alan McKinnon wrote: Apparently, though unproven, at 20:54 on Friday 31 December 2010, walt did opine thusly: So, if you've ever run the NVIDIA install program manually, you probably have two different nvidia.ko files in /lib/modules, and the wrong one gets loaded automatically at boot time. Re-emerging nvidia-drivers will build *and* load the correct kernel module each time you do it. May not be your problem but it's easy to check, at least. I didn't think I'd ever run the nVidia installation program, and on checking I see one nvidia.ko for each kernel version I have installed, so that isn't the problem. module-rebuild takes care of all that nicely. And it wants to rebuild the module again, even though I only remerged it yesterday. Something is messing about with the nVidia module. That's how module-rebuild works, it's not broken. It *will* rebuild everything. It can't rely on the normal version numbers and USE flags to know if something needs updating, as the problem it is designed to solve is when you do a kernel upgrade and leave yourself without the out-of-tree modules you need in that new kernel version. If you use nvidia version X, then you need the version X kernel modules in every /lib/modules/kernel/ you use. So it just rebuilds everything every time. -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] disk /dev entry
Apparently, though unproven, at 23:38 on Saturday 01 January 2011, Daniel D Jones did opine thusly: This is more of a curiosity question than a problem. I just added a new diskdrive to my system. It's the same model as one I already have installed. lshw shows the following for the two disks: *-disk:2 description: ATA Disk product: ST31000528AS vendor: Seagate physical id: 0.0.0 bus info: s...@2:0.0.0 logical name: /dev/sdc version: CC37 serial: 9VP21EZB size: 931GiB (1TB) capabilities: partitioned partitioned:dos configuration: ansiversion=5 signature=2adeb97b *-disk description: ATA Disk product: ST31000528AS vendor: Seagate physical id: 0 bus info: i...@1.0 logical name: /dev/hdc version: CC3E serial: 9VP9G4VW size: 931GiB (1TB) capabilities: ata dma lba iordy smart security pm partitioned partitioned:dos configuration: signature=1e89b64b smart=on These are both SATA disks. Why is the first one showing up as /dev/sdc and the second as /dev/hdc? I thought all SATA disks would show up as sd* and EIDE disks showed up as hd*. I note that the bus info is different - one showing the scsi bus and one the ide bus. Both devices are plugged into a row of SATA ports on my mobo. Maybe the new disk is set to IDE mode as shipped? -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] disk /dev entry
On 1/1/2011, at 9:38pm, Daniel D Jones wrote: These are both SATA disks. ... I thought all SATA disks would show up as sd* and EIDE disks showed up as hd*. That's weird. In `make menuconfig` what does this part say? Device Drivers --- ATA/ATAPI/MFM/RLL support (DEPRECATED) * Serial ATA and Parallel ATA drivers --- Stroller.
[gentoo-user] New project in perl? {OT}
I'm sorry this is OT but I really value the opinion of many people subscribed to this list. I'm starting a new project that is quite straightforward and will interface with an old project. The only point of contact between the two projects might be both of them having access to the same database table. The old project is written in a language that is related to perl so I can imagine there would be some benefit to using perl for the new project. Am I foolish to start a new project in perl at this stage in its lifecycle? I won't be doing the coding myself and I wonder if I would be better off with PHP since more coders seem to be familiar with PHP than perl. - Grant
Re: [gentoo-user] Re: nvidia update problems
On Saturday 01 January 2011 20:26:11 Alan McKinnon wrote: Apparently, though unproven, at 12:36 on Saturday 01 January 2011, Peter Humphrey did opine thusly: Something is messing about with the nVidia module. That's how module-rebuild works, it's not broken. It *will* rebuild everything. Yes, I wasn't impugning module-rebuild. All the same, something does appear to be messing about with the nVidia module. -- Rgds Peter. Linux Counter 5290, 1994-04-23.
Re: [gentoo-user] New project in perl? {OT}
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 02/01/2011, at 09:04, Grant wrote: I'm sorry this is OT but I really value the opinion of many people subscribed to this list. I'm starting a new project that is quite straightforward and will interface with an old project. The only point of contact between the two projects might be both of them having access to the same database table. The old project is written in a language that is related to perl so I can imagine there would be some benefit to using perl for the new project. Am I foolish to start a new project in perl at this stage in its lifecycle? I won't be doing the coding myself and I wonder if I would be better off with PHP since more coders seem to be familiar with PHP than perl. TBH use neither, most people are jumping away from PHP and Perl. There is no issue with a change to your language now. SQL is a standard so using python, or ruby to interact with it will have no issues. Just make sure that you copy the database to a dev box first so that you avoid mangling your important data. - Grant William Brown pgp.mit.edu -BEGIN PGP SIGNATURE- Version: GnuPG/MacGPG2 v2.0.16 (Darwin) iQIcBAEBAgAGBQJNH8QuAAoJEHF16AnLoz6Jj3kP/Rb+2kRj3CYzfp9sO3EgTnfw b7SHcz/wJaHPyK+fHftWJRAfpa7+8ymNmSLuQj/9/lBnn5/7OOw0GhGLC1zHJ96u nQLdToRXreSDd4ci6k9uVdlNP9qjrSSjBG3jAP5ZWnDwq7vNhAPD4M67i03uLCYC B1hJv+ZtFk98U6HivkJv9wO6GoE/QRNVtX1BS6y6ZvjiRF3qx5PPDRxAmqfgbDzk E04xBUOsXH1+yJKewZdoM3bIHUNxYpDK9IaSjDQVPJK8/TFn7ImNRP0aDbuUIpuH b95Ujq3ugbmE4ZD8hysD6oIgc8iGnEpmGuBCWuA3a/1VVx+e1p+qyPSriDVeFh15 gWKuxS6lAlqwlpXKrqnKMTfKzHivPTkw2/muMKAB09nC8lVEUvrj+K2es3rFfUI8 ZbiBIAVSJThgFLJfbWOQoh6AB6qqP5BMCAYtVTuuzKDVtc6ww5lipX4dejRooDl/ P9uAF1Yv8q9X3DNIF3LjVpPxEjzqUMFCKIUXQ1BDCUhST7YNPAyjxbbf5gTEcDN8 7XVBcCaT/fjtF/gxq2KKP/jlJLwgzuMxsmxON9snK8rUgODfGz1yutZOH70u7Gmg irC4V8uE1lNZVJgl/MQ2qXIBwcEtH+/+5eRvMHxSwtigNUqZcr0Re1h8Csk/rSuX XFGWRO4tas3sH+ReOB0e =QxHa -END PGP SIGNATURE-
Re: [gentoo-user] Core i7 M620 power management problem
On Saturday 01 January 2011 23:50:21 Bill Longman wrote: On 01/01/2011 03:16 AM, Stefan G. Weichinger wrote: Am 2010-12-31 11:59, schrieb Mick: Hmm ... could it be a buggy BIOS? Are you running the latest firmware for it? Yes, that would also have been my next question. Maybe you even *find* a bug in that BIOS right now that should be corrected. Well the good news is that my BIOS is not foobarred. The bad news is that I have to figure out what's wrong -- Ubuntu 10.10 came up with scaling_max_freq = 2667000. ARGH Changing some kernel settings Did you diff the kernel configs to see what's different between the two OS'? -- Regards, Mick signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] New project in perl? {OT}
On 1/1/2011 2:34 PM, Grant wrote: I'm sorry this is OT but I really value the opinion of many people subscribed to this list. I'm starting a new project that is quite straightforward and will interface with an old project. The only point of contact between the two projects might be both of them having access to the same database table. The old project is written in a language that is related to perl so I can imagine there would be some benefit to using perl for the new project. Am I foolish to start a new project in perl at this stage in its lifecycle? I won't be doing the coding myself and I wonder if I would be better off with PHP since more coders seem to be familiar with PHP than perl. In '99 I worked with a fellow who styled himself a software architect. The first step of each project he managed involved stating We will write this software in Java. As you can imagine that's sorta backwards. I'd spec the software function, features, etc and then decide which language has better tools or command of the problem space. You will have to balance that against your knowledge of the language and the developer skills you have access to. However even the exercise of deciding Python appears to be the superior language in this problem space, but we're going to go with Perl because the database module for our db already exists and is much more mature. Bob knows Perl better too. is worth doing because it helps define the scope of the project. FWIW the current startup I'm at is using Ruby for the front end and it's been a bit more work that PHP which is what the last company used. That's partly Rails immaturity, our lack of experience with Ruby, and having to learn the Rails/Ruby way. Unless the language you're familiar with is completely unsuitable, I'd say familiarity trumps language features. YMMV. kashani
Re: [gentoo-user] New project in perl? {OT}
On 1/1/2011, at 10:34pm, Grant wrote: ... I'm starting a new project that is quite straightforward and will interface with an old project. The only point of contact between the two projects might be both of them having access to the same database table. The old project is written in a language that is related to perl so I can imagine there would be some benefit to using perl for the new project. Am I foolish to start a new project in perl at this stage in its lifecycle? I won't be doing the coding myself and I wonder if I would be better off with PHP since more coders seem to be familiar with PHP than perl. I'm not sure if I've mentioned before, but I picked up Perl fairly recently (within the last 12.5 months) although I haven't done *that* much with it. I *really* like Perl. It feels extremely robust and right. I originally picked up Perl in order to parse the output of another program and build an HTML table based on that data. The other program happened to be written in Perl, too, but I figured that Perl was a good choice because it was supposed to be good at parsing (and parsing was the job I was trying to do). I started out parsing this output, and it turned out that the program I was depending on didn't give enough information externally this way. So I had to modify the original program, and add the feature I needed - I was able to do so really fairly quickly, and soon had my first submission (a patch of maybe 200 lines) accepted into an open-source project. Well, I guess vgetty distributes a shell script of mine in its contrib directory, but this felt much more of a grown-up achievement. Doing this in Perl felt really accessible to me, being able to complete this task within hours [1] of picking up the language. The stage of Perl's lifecycle is not anything to be worried about. Perl may not be a cool language, but it isn't going away. Perl 5.x.y will be maintained for a long time; Perl 6 is in development. On the other hand, you can easily find Perl developers who have been using the language in industry for a decade. You can write bad code in any language - Perl has a reputation for opacity but, y'know, Greek sounds pretty opaque to me, that doesn't mean it's a bad language. To be honest, I think the problem is that Perl can be really terse, and that's why newcomers to a codebase have problems understanding it, but I'm inclined to think of Perl's terseness as a *good* thing. From the code I've read on the net, the advice I've received from Perl programmers, I tend to feel the average code quality of Perl developers is higher than that of the average coder in some more fashionable languages, such as PHP or Python. I would think that if you were to contact your local PM group you would find someone who writes pretty good code and who will provide references. http://www.pm.org/groups/ Perl has a big library of modules for interfacing with databases. Your guy should use as many of those as he needs to, and not reimplement things from scratch. Beware of people who write their own libraries or who prefer to do it for themselves - there are some extremely sophisticated and well-maintained modules in CPAN. Stroller. [1] Not the same day, but within only a few hours of actual using the language, for large values of a few.
Re: [gentoo-user] Good file system that recovers from a power failure.
On Fri, 2010-12-31 at 19:12 -0600, Dale wrote: Alan McKinnon wrote: Apparently, though unproven, at 00:27 on Saturday 01 January 2011, Dale did opine thusly: It's my opinion that reiser is in security-fix-only mode from whoever is maintaining it. If everything else around it stays the same, the fs will obviously continue working just as it always did. But the surrounding system is not stable, it changes rapidly, especially in kernel space, so the odds are stacked against reiser for bitrot. For all these reasons, I regretfully switched my own systems over to ext4 some time ago. Rieser was a good fs whose time has come and gone and I no longer had warm and fuzzies about the future with it. I'm not sure I EVER saw a update to reiserfs. I was hoping it was just that good. lol This is also the reason I was considering moving to ext4 or something. How has ext4 been treating you since the switch? I also assume you have UPSs as well? It's still early days, but ext4 has been good here on all machines. I don't have a UPS (couldn't be bothered really...) so the UPS is the device's battery. Which means me doing something really stupid and locking the machine up is the most common reason for hard reboots. It survived every time so far. That sounds good. Your situation is not a theory but real and in practice. We all know what happens to theories. :-( Dale :-) :-) As someone who suffered from a rare, but fatal bug in some kernel updates to reiserfs3 which were only fixed in a recent versions, I can say yes, its still being actively maintained. But, I am thinking its time to move on - in particular something that can fsck online (my mythtv and backup archives on reiserfs have to be done offline and its takes hours to do terrabytes!) and still be robust. I use dirvish for backups which creates a LOT of hardlinks which can be very hard on a file system. ext2 typically lasts only a few cycles, while ext3 is only a little better even with full journalling. Coupled to the fact neither is very good with power cuts and they are a worst case choice for data security :) Reiserfs3 by contrast is very very good, with only a few instances of problems over many years (since beore 3 was even in the kernel) - none of which have lost critical data or file systems (ext2/3 devs, are you listening :) Even the slowpath bug I ran into just required a kernel downgrade and an fsck until later kernel versions fixed the bug. I am now trying btrfs and am very impressed. On line fsck is wonderful and I have had one instace of corruption due to a flakey hard disk - the partition is on lvm so I moved it to another disk in the array and its been solid since - didnt lose it or any any data. The dirvish backups are fast enough (impression only, no timings) but large scale deleted (60Gb copies of laptops etc) are much slower than reiserfs. My only glitch has been dirvish/btrfs inability to deal with a .gvfs file in some home directories - it has some wierd permissions but an exclusion from the backup regime bypasses it. reiserfs doesn't have a problem with it. So, for me at least, btrfs is looking like the way forward. Its in testing at the moment, but I am ready to move whole systems over to it. BillK -- William Kenworthy bi...@iinet.net.au Home in Perth!
Re: [gentoo-user] Core i7 M620 power management problem
On Sat, Jan 1, 2011 at 4:34 PM, Mick michaelkintz...@gmail.com wrote: Did you diff the kernel configs to see what's different between the two OS'? There was no /proc/config.gz. How do you find it without that? I looked through the proc tree but didn't find anything. I added some printk's to the kernel and I see some segfaults. I'll try another kernel tomorrow and see if tuxonice gives me anything different. -- Bill Longman
Re: [gentoo-user] Good file system that recovers from a power failure.
On Sat, Jan 01, 2011 at 03:09:24AM -0500, Philip Webb wrote Total corruptions in 5 years with reiserfs-3.6 and NO ups in that environment = zero. Same here: I've used Reiserfs since 2003 never had a problem; AOL Me too /AOL. Reiserfs 3 for several years and no lost data. -- Walter Dnes waltd...@waltdnes.org