On 7/6/06, Scott R. Godin <scott.g@xxxxxxxx> wrote:
On Thu, 2006-07-06 at 22:36 +0800, Deepak Shrestha wrote: > I guess there are also other factors behind the corrpution of > directories and other problems (may be bug in fedora itself). but so > far i haven't seen any weiredness in my box yet as mentioned before > like directory corruptions. Its working well and I am happy. > no, that's the thing. it doesn't show up right away. everything runs fine for weeks to months.. and then BOOM What I'd HIGHLY recommend to you is this: use tune2fs (manpage has details) to assign parameters to your various partitions so that they get checked at boot time automatically regularly. Why? well, the manpage points out: "You should strongly consider the consequences of disabling mount-count-dependent checking entirely. Bad disk drives, cables, memory, and kernel bugs could all corrupt a filesystem without marking the filesystem dirty or in error. If you are using journaling on your filesystem, your filesystem will never be marked dirty, so it will not normally be checked. A filesystem error detected by the kernel will still force an fsck on the next reboot, but it may already be too late to prevent data loss at that point." I, being paranoid, set mine thusly, for example (for my / partition among others): tune2fs -c 25 /dev/sda3 tune2fs -i 3w /dev/sda3 so it gets automatically checked at boot time every 25 mounts or at most 3 weeks, whichever comes first. I did the same or similar with the rest of my partitions, excluding only the non-ext3 (windows, swap, etc.) partitions. If you never reboot, it'll never get checked, though. However if you experience a crash, this will help cover your ass by running fsck on the system if you're beyond the limit. additionally, I set up a crontab entry to run rpm -Va once a week and report to me when things have changed. (I've already set up /etc/aliases to forward root's mail to my user account, and ran newaliases to update the db) (su -) to root (making sure to use '-' .. see man su for details why), and then crontab -e, and enter the following: 5 5 * * sun rpm -Va 2>&1 meaning: Run: rpm -Va 2>&1 At: 5:05am every Sunday I'm still picking up the pieces.. today I discovered that /usr/bin/cpan had turned into a socket .. I still haven't replaced all the corrupted files with fresh installs of their respective rpms, but am slowly getting there a little bit at a time in between the work I'm doing. good luck with it, but keep your eyes peeled for weirdness.
Don't you think its a tad bit odd, that no one else is reporting this mess? Look at the nvnews forums and see how many people are reporting filesystem corruption. You wont' find any, and there are alot of people reporting problems over there. You stated that you ran memtest86, and it didn't find any problems. Did you check all the rest of your hardware? Maybe your CPU is bad? Maybe your disk(s) is failing. Maybe your system is overheating? Maybe there was a spike in your electricity? Screaming fire in a crowded theater when your pants are on fire seems a bit ridiculous. -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ L. Friedman netllama@xxxxxxxxx LlamaLand http://netllama.linux-sxs.org
