>>> On Thu, 15 Sep 2005 08:38:16 -0500, Dave Kleikamp >>> <[EMAIL PROTECTED]> said:
shaggy> I'm kind of buried in work right now, but your problems shaggy> appear serious, so I'm trying to respond as quick as I shaggy> can. Not that serious, things mostly work (and fairly well), I just get the rather occasional crash. Real corruption would be more serious. No need to hurry, it is also just my home PC :-). >> * Some of my tests were tree traversals, [ ... ] the timings >> be with '-o noatime', unfortunately I got a crash because >> of that. shaggy> noatime is used a lot, I don't think noatime was a shaggy> direct cause of the problem, [ ... ] I concur. I have tried to reproduce it subsequently but did not happen again, but then I now remember I did this on an empty filesystem being freshly loaded, so perhaps it is a corner case. I'll do another test. >> * When converting from 'ext3' to JFS file systems, I did this >> by copying things around, and I got a couple of lockups. It >> may be that these were related to high buffer cache traffic >> (I was doing a large 'dd' between partitions at one time) >> and races thereof. shaggy> No idea what could be happening here. If you could shaggy> capture a stack trace of the processes, it may give me a shaggy> clue what's going on. 'echo t > /proc/sysrq-trigger' shaggy> should dump the stack traces to the syslog. Looks like a semi-hard lockup, but I will try to see what I can do, and perhaps run a kernel debugger. Semi-hard means I can ping the machine from a laptop, but nothing else works, and the screen and keyboard are dead. [ ... ] shaggy> I had another recent report of dtree corruption [ ... ] shaggy> (I highly doubt that you ran mkfs -O to create the shaggy> partition.) I didn't so it should not be that. BTW, that '-O' is a really useful and just about unique feature of JFS, and perhaps it should be mentioned more prominently. http://OSS.SGI.com/projects/xfs/papers/ukuug2003.pdf «FUTURE * Case insensitive support. Big speedup for Samba» http://lists.Samba.org/archive/samba/2004-August/090489.html «> I'm wondering what issues / problems I might run into if > I'm using samba 3 on linux with the IBM JFS filesystem > installed to be case-insensitive. I've had no luck > finding anyone who says they've tried it. It should be really fast :-). Just set case sensitive = yes in the [global] section and let the filesystem sort it out :-).» >> * Making a file system with a 30MiB log instead of the >> default 32MiB makes reading it with 'tar' over twice as >> slow. This for the same partition on the same hard disc >> with the same content freshly loaded (it was so strange I >> checked several times). shaggy> You're right that this is strange. If you are running shaggy> with noatime, the journal shouldn't be a factor at all shaggy> when reading the volume. This one really puzzles me. But I stopped using 'noatime' after the first attempt, as it had caused trouble. BTW I should add I used 'jfs_debugfs' 'su p' and the output was nearly identical, that is except for the size and related values for the log. I'll try to reproduce it. [ ... dtree trouble ... ] shaggy> If you see this again, run jfs_fsck with the -v flag. shaggy> That may give me a better idea of the nature of the shaggy> dtree corruption. OK, that sounds good. [ ... ] shaggy> I think DEBUG_PAGEALLOC helped uncover this bug. Ha, as to this I have almost all kernel selfchecking options enabled all the time. The slowdown is not large and I feel a lot better even if it is just my home PC :-). Just about the only selfcheck I have disabled is the slab allocator selfchecks, because while in 2.4 they did not cost a lot, in 2.6 they are very slow. Oh well. Would it help and not be too intrusive for my daily use if I enabled the JFS debugging? ------------------------------------------------------- SF.Net email is sponsored by: Tame your development challenges with Apache's Geronimo App Server. Download it for free - -and be entered to win a 42" plasma tv or your very own Sony(tm)PSP. Click here to play: http://sourceforge.net/geronimo.php _______________________________________________ Jfs-discussion mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/jfs-discussion
