(Illustration by Gaich Muramatsu)
On Fri, Jul 16, 2004 at 02:10:03PM -0400, Jan Harkes wrote: > On Fri, Jul 16, 2004 at 11:19:50AM -0500, Troy Benjegerdes wrote: > > This was with a ppc client, and a replicated volume on two ppc servers. > > (coda 6.0.6 release debian packages rebuilt for ppc) > > > > I was attempting to copy my maildir into a coda volume. At the same > > time, I had my laptop untarring the linux kernel into a different > > volume. > > > > 03:01:16 Reintegrate: u.hozer.mail, 100/1007 records, result = SUCCESS > > 03:01:18 fatal error -- fsobj::dir_Create: > > (1087644725.M331486P22611V0000000000000900I0011D52C_1.kalmia.hozed.org,S=2832, > > 6f558048.7f000005.cbf0.339d) Create failed! > > 03:01:23 RecovTerminate: dirty shutdown (1 uncommitted transactions) > > > > Is there something I should look for in the server logs as to why this > > failed? > > It looks like the maildir copy failed here. What is the size of the > directory you were copying? > > Oh, and it definitely looks like your servers are unable to resolve > their differences, so their contents will slowly keep diverging until > everything is considered a conflict, or their respective resolution > logs are filled up. On my linux-kernel torture test.. there seem to be some issues when the quota fills up.. (but it is resolving a bunch of stuff) 15:24:33 RS_ForceFile: Error 122 in AdjustDiskUsage 15:24:33 Entering VFlushVnode for vnode 0x2f9c 15:24:33 GetAttrPlusSHA: Computing SHA 29000003.2f9c.3498, disk.inode=0 15:24:34 GetAttrPlusSHA: Computing SHA 29000003.2ec0.3461, disk.inode=0 15:24:35 Volume 687865859 (testrep.0) is full 15:24:35 RS_ForceFile: Error 122 in AdjustDiskUsage 15:24:35 Entering VFlushVnode for vnode 0x2ec0 15:24:35 GetAttrPlusSHA: Computing SHA 29000003.2ec0.3461, disk.inode=0 15:24:36 GetAttrPlusSHA: Computing SHA 29000003.32c6.1a4a, disk.inode=0 What happens if I just 'rm -rf' the kernel tree? Will there be orphans that get left over? Or will it (in theory) all get cleaned up?Received on 2004-07-16 16:32:36