Coda File System

Re: Error on check-point/hoarding

From: Jan Harkes <jaharkes_at_cs.cmu.edu>
Date: Mon, 1 May 2000 13:32:11 -0400
On Mon, May 01, 2000 at 08:58:55AM -0700, Tom Tarka wrote:
> just after completing a big copy into a coda volume (85M),
> venus died, saying the the server was dead (while it's not,
> although it did give an error) and then died after trying to
> hoard/walk.
> 
> ideas on what went wrong?  I can reproduce with higher
> debugging levels as needed...
> 
...
> 09:27:23 StoreBulkTransfer: CheckSE failed (-2014), (64000001.148.145)

That error is SE_FAIL2, which is used all over the place for fatal
errors. f.i. The server believes it is on the client side of the SFTP
transfer. Sort of like a panic, don't know what triggered it.

> [ T(01) : 2686 : 08:47:48 ] cache::EndRvmTruncate
> 08:48:13 Fatal Signal (11); pid 1834 becoming a zombie...
> 08:48:13 You may use gdb to attach to 1834

Hmm, only attaching gdb to the client and grabbing a backtrace can tell
us where we crashed here.

Jan
Received on 2000-05-01 13:47:00