I had a 2008.11 machine crash while moving a 700gb file from one machine to another using cp. I looked for an existing bug for this, but found nothing. Has anyone else seen behavior like this? I wanted to check before filing a bug. cheers, Blake
On Thu, 5 Mar 2009, Blake wrote:> I had a 2008.11 machine crash while moving a 700gb file from one machine > to another using cp. I looked for an existing bug for this, but found > nothing. > > Has anyone else seen behavior like this? I wanted to check before > filing a bug.Have you got a copy of the stack dump? That would make it easier to track down. Regards, markm
I have savecore enabled, but it doesn''t look like the machine is dumping core as it should - that is, I don''t think it''s a panic - I suspect interrupt handling. Speaking of which, does OpenSolaris support Plug''n''Play IRQ assignment? On Thu, Mar 5, 2009 at 3:12 PM, Mark J Musante <mmusante at east.sun.com> wrote:> On Thu, 5 Mar 2009, Blake wrote: > >> I had a 2008.11 machine crash while moving a 700gb file from one machine >> to another using cp. ?I looked for an existing bug for this, but found >> nothing. >> >> Has anyone else seen behavior like this? ?I wanted to check before filing >> a bug. > > Have you got a copy of the stack dump? ?That would make it easier to track > down. > > > Regards, > markm >
On Fri, 6 Mar 2009, Blake wrote:> I have savecore enabled, but it doesn''t look like the machine is dumping > core as it should - that is, I don''t think it''s a panic - I suspect > interrupt handling.Then when you say you had a machine crash, what did you mean? Did you look in /var/crash/* to see if there''s something there? If not, it''s possible your dump zvol just needs to have its core retrieved, which means you can just run ''savecore'' at any time and get it. Regards, markm
I have savecore enabled, but nothing in /var/crash: root at filer:~# savecore -v savecore: dump already processed root at filer:~# ls /var/crash/filer/ root at filer:~# On Fri, Mar 6, 2009 at 4:21 PM, Mark J Musante <mmusante at east.sun.com> wrote:> On Fri, 6 Mar 2009, Blake wrote: > >> I have savecore enabled, but it doesn''t look like the machine is dumping >> core as it should - that is, I don''t think it''s a panic - I suspect >> interrupt handling. > > Then when you say you had a machine crash, what did you mean? > > Did you look in /var/crash/* to see if there''s something there? ?If not, > it''s possible your dump zvol just needs to have its core retrieved, which > means you can just run ''savecore'' at any time and get it. > > > Regards, > markm >
On Fri, 6 Mar 2009, Blake wrote:> I have savecore enabled, but nothing in /var/crash: > > root at filer:~# savecore -v > savecore: dump already processed > root at filer:~# ls /var/crash/filer/ > root at filer:~#OK, just to ask the dumb questions: is dumpadm configured for /var/crash/filer? Is the dump zvol big enough? How do you know the whole machine crashed instead of, for example, the X server just coincidentally died? I''m not trying to be difficult, just trying to narrow down where the problem is. If you file a bug that just says ''cp killed my machine'', it''s likely to be marked incomplete until there''s a reliable way of reproducing it. Regards, markm
These are fair questions, answered inline below :) On Fri, Mar 6, 2009 at 4:45 PM, Mark J Musante <mmusante at east.sun.com> wrote:> On Fri, 6 Mar 2009, Blake wrote: > OK, just to ask the dumb questions: is dumpadm configured for > /var/crash/filer? ?Is the dump zvol big enough? ?How do you know the whole > machine crashed instead of, for example, the X server just coincidentally > died?root at filer:~# dumpadm Dump content: kernel pages Dump device: /dev/zvol/dsk/rpool/dump (dedicated) Savecore directory: /var/crash/filer Savecore enabled: yes I walked over to the console to see the machine rebooting. gdm is disabled on this box. One point of possible interest is that the target fs is on a zpool with ''compression=6'' enabled - that is, gzip level 6 compression. The machine has 4gb ram, which means the 2gb dump zvol should be big enough, IIRC.> > I''m not trying to be difficult, just trying to narrow down where the problem > is. ?If you file a bug that just says ''cp killed my machine'', it''s likely to > be marked incomplete until there''s a reliable way of reproducing it. > > > Regards, > markm >