also sprach Adorjáni Gábor <adi@adi.priv.hu> [2006.02.01.2320 +0100]: > Don't even think about using XFS, if you plan to use even a moderately > loaded MySQL instance on top of it. It locked up one of our servers I wouldn't want this to become a flamefest, but thankfully, I would never have a MySQL instance on an XFS filesystem, or any filesystem for that matter. :) I can confirm that PostgreSQL, even very loaded (we have an instance here that occupies 71 Gb), works like a charm on an XFS filesystem. > several times to the point where only the service processor's > watchdog function helped, and resetted the machine. There were no > signs, no logs, it simply blocked all disc I/O. The system is > a Debian Sarge with a custom 2.4.31 kernel, but according to the > info I found via Google, the problem is still in the 2.6 series. I hope you have reason to believe that it's XFS at fault, not MySQL. Sure, XFS is not bug-free, but neither is MySQL. If the problem doesn't appear when the instance is on another FS, I suppose there's an argument. However, as computers go, you're probably going to be unable to reproduce the situation with another filesystem, so it's going to be impossible to say it was XFS' fault. Anyway, I'd be interested (off-list if you deem it inappropriate for this thread) to learn about the problem you found that allegedly still exists in the 2.6 series. -- Please do not send copies of list mail to me; I read the list! .''`. martin f. krafft <madduck@debian.org> : :' : proud Debian developer and author: http://debiansystem.info `. `'` `- Debian - when you have better things to do than fixing a system Invalid/expired PGP (sub)keys? Use subkeys.pgp.net as keyserver! beauty, brains, availability, personality; pick any two.
Attachment:
signature.asc
Description: Digital signature (GPG/PGP)