I had a machine die last night at around 2:01am. That does *not* correspond with any cron jobs. When I got remote hands in front of it, this message was streaming on the console (copied from what remote hands said verbally, the formatting may be off): swap_pager: indefinite link buffer device:#idad/0x20001,blkno:280,size:4096 Every console message said the same thing. There are no errors in the system's log files. The hardware is a dual PII 450Mhz Compaq 1850R with the Compaq SMART-2SL array controller with a 3x4.5GB RAID 5 configuration. It uses the ida controller. I have 5 other identical boxes in the same configuration, doing different jobs, that have been working well for the past several months. This is the lowest loaded of my 1850R boxes and has been working just fine since it was put in production two months ago. FreeBSD 4.8-RELEASE-p7 #0: Thu Sep 18 15:06:09 EDT 2003 (huh, wonder how that one got left behind.) runs ipfw, ssh, postfix, mailman, and apache. I see messages in the archives about: swap_pager: indefinite wait buffer: blah blah But nothing that matches the "indefinite link buffer" message. And I just lost access to the box again. This has a high probability of being a hardware issue but I want to run it past the group just in case it rings any bells for anyone. -- Scott Lambert KC5MLE Unix SysAdmin lambert@lambertfam.org
On Sat, 8 Nov 2003, Scott Lambert wrote:> I had a machine die last night at around 2:01am. That does *not* > correspond with any cron jobs. > > When I got remote hands in front of it, this message was streaming on > the console (copied from what remote hands said verbally, the formatting > may be off): > > swap_pager: indefinite link buffer device:#idad/0x20001,blkno:280,size:4096Your swap device went to lunch. (Usually the message is "indefinite wait buffer" but I haven't checked.)> Every console message said the same thing. > > There are no errors in the system's log files.Probably since the rdrive that logs go to went to lunch too.> The hardware is a dual PII 450Mhz Compaq 1850R with the Compaq SMART-2SL > array controller with a 3x4.5GB RAID 5 configuration. It uses the ida > controller. > > I have 5 other identical boxes in the same configuration, doing > different jobs, that have been working well for the past several months. > This is the lowest loaded of my 1850R boxes and has been working just > fine since it was put in production two months ago.This is probably like a flaking dpt we have at work. I flashed the BIOS on it but will try replacing the cable next. Something is failing and not picking up the problem. I wonder if you could try to provoke it by doing a rebuild or verify on the volume(s) on the controller. Also try uupdating the firmware for kicks, if its not up to date. -- Doug White | FreeBSD: The Power to Serve dwhite@gumbysoft.com | www.FreeBSD.org