Frankly, this deserves a higher bug rating than "important". A Unix without a working NFS is not a Unix. At least not a usable one.
I can now confirm that downgrading to 2.6.18 seems to fix it. All 3 machines I have with diskless Lenny are still up and usable. By this time they would have been out of commission with 2.6.26. On Wed, 2009-04-15 at 13:52 -0400, John Morrissey wrote: > On Wed, Apr 15, 2009 at 01:30:51PM +0100, Anton Ivanov wrote: > > Overtime any system with NFS usage grinds to a crawl. While root or /usr > > on NFS are most affected the same problem should affect other NFS usage. > > The symptoms are extremely high system load, > [snip] > > The reason seems to be: > > http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git&a=commitdiff&h=23918b03060f6e572168fdde1798a905679d2e06 > > FWIW, we're also seeing this, on machines mounting several large NFS volumes > that serve virtual accounts (i.e., they all have the same UID/GID and no > shell access). > > After ~two days of uptime, these moderately loaded machines (running Apache, > ProFTPD, and Courier IMAP) start spending >90% of CPU time in system state > and become so unresponsive that they must be rebooted. Running the 2.6.28 > that was recently in sid (which contains the commit Anton mentions) fixed > this. > > john -- Understanding is a three-edged sword: your side, their side, and the truth. --Kosh Naranek A. R. Ivanov E-mail: aiva...@sigsegv.cx WWW: http://www.sigsegv.cx/ pub 1024D/DDE5E715 2002-03-03 Anton R. Ivanov <ariva...@sigsegv.cx> Fingerprint: C824 CBD7 EE4B D7F8 5331 89D5 FCDA 572E DDE5 E715 -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
