Date: Thu, 11 May 1995 11:59:06 +0800 (CST) From: Brian Tao <taob@gate.sinica.edu.tw> To: FREEBSD-HACKERS-L <freebsd-hackers@FreeBSD.org> Subject: Lockup on heavy NFS client? Message-ID: <Pine.BSI.3.91.950511114334.197A-100000@aries.ibms.sinica.edu.tw>
next in thread | raw e-mail | index | archive | help
I ran my HTTP benchmark again last night, but this time with the
test documents (1900 files in 10 directories, 53 megabytes) NFS-
mounted from a local Ethernet 486DX2/66 running FreeBSD 2.0-950322.
All seemed fine for the first hour, so I left for home. When I came
back in this morning, my DX4 (which was the httpd server and the NFS
client) was hung, with a blank display. I couldn't get to a console
display to see what had caused the crash. X was not running at the
time. Ctrl-Alt-Del was dead too. Reboot time.
Meanwhile, the DX2 (the NFS server) was still humming along, but
with these messages on console:
May 11 03:04:25 virgo /kernel: nfs server aries:/usr/share: not responding
May 11 03:40:23 virgo /kernel: nfs server aries:/usr/share: not responding
May 11 04:10:24 virgo /kernel: nfs server aries:/usr/share: not responding
May 11 04:38:08 virgo /kernel: nfs server aries:/usr/local: not responding
May 11 04:38:11 virgo /kernel: vnode_pager_input: I/O read error
May 11 04:38:11 virgo /kernel: vm_fault: pager input (probably hardware) error, PID 1003 failure
May 11 04:40:33 virgo /kernel: nfs server aries:/usr/share: not responding
May 11 04:51:20 virgo /kernel: nfs server aries:/usr/local: not responding
May 11 04:51:21 virgo /kernel: vnode_pager_input: I/O read error
May 11 04:51:21 virgo /kernel: vm_fault: pager input (probably hardware) error, PID 1003 failure
May 11 05:04:30 virgo /kernel: nfs server aries:/usr/local: not responding
May 11 05:04:31 virgo /kernel: vnode_pager_input: I/O read error
May 11 05:04:31 virgo /kernel: vm_fault: pager input (probably hardware) error, PID 1003 failure
May 11 05:17:40 virgo /kernel: nfs server aries:/usr/local: not responding
May 11 05:17:41 virgo /kernel: vnode_pager_input: I/O read error
May 11 05:17:41 virgo /kernel: vm_fault: pager input (probably hardware) error, PID 1003 failure
May 11 05:30:50 virgo /kernel: nfs server aries:/usr/local: not responding
May 11 05:30:51 virgo /kernel: vnode_pager_input: I/O read error
[etc]
This continued like clockwork, every 13 minutes 10 seconds (780
seconds) until I decided to reboot it too. No other messages
indicating a problem on the NFS server.
Back on the DX4, nothing in the syslog, access_log or error_log
files would indicate anything that would cause the machine to stop.
The last HTTP access was at 2:47 am (damn, and I had just left 15
minutes earlier too...) and the benchmark had been running for nearly
an hour by then.
Anyone seen this problem before? I can provide more details if
necessary. And what's with the 780-second cycle?
--
Brian ("Though this be madness, yet there is method in't") Tao
taob@gate.sinica.edu.tw <-- work ........ play --> taob@io.org
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSI.3.91.950511114334.197A-100000>