On Fri, Oct 24, 2003 at 01:30:18PM +0200, Poul-Henning Kamp wrote: > You may want to try this patch, Kirk sent it to me in response to a > case where my NFS-server hung during snapshot processing. > > I have only just managed to put it on my server so I don't know if > it solves the problem or not. Hmm, my NFS server stays ok; it's the client that has problems. Until somewhere this week, the following occurred: prompt> portupgrade -rR wv-0.7.5 ---> Upgrading 'imake-4.3.0' to 'imake-4.3.0_1' (devel/imake-4) ---> Building '/usr/ports/devel/imake-4' ===> Cleaning for perl-5.6.1_14 ===> Cleaning for imake-4.3.0_1 ===> Extracting for imake-4.3.0_1 >> Checksum OK for xc/X430src-1.tgz. >> Checksum OK for xc/X430src-3.tgz. [... ctrl-t (stty status) (15 seconds interval):] load: 2.43 cmd: tar 12295 [nfsfsync] 0.09u 4.16s 2% 544k load: 2.40 cmd: tar 12295 [nfsfsync] 0.09u 4.16s 2% 544k load: 2.34 cmd: tar 12295 [nfsfsync] 0.09u 4.16s 2% 544k load: 2.17 cmd: tar 12295 [nfsfsync] 0.09u 4.16s 2% 544k load: 2.05 cmd: tar 12295 [nfsfsync] 0.09u 4.16s 2% 544k load: 2.00 cmd: tar 12295 [nfsfsync] 0.09u 4.16s 2% 544k In other words, the 'tar' is dead in the water. Since somewhere this week, it not only hangs up that specific process, but all processes that try using NFS hang themselves up and eventually the machine reboots... I haven't produced a crash dump yet. The NFS server I use runs FreeBSD 4 and is mounted by a lot of FreeBSD 4 machines that do not have this problem. ZloReceived on Fri Oct 24 2003 - 03:14:07 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:26 UTC