[OpenAFS] Re: Help with wedged Solaris box

[Kevin Hildebrand]

I just checked another crash dump and it looks the same, no 
afs_CacheTruncateDaemon.  The currently running machine _does_ have one,
so it's dying somewhere.  I don't see anything in the system logs anywhere 
that indicate anything unusual.

More information from the users of the machine indicate that the hangs 
appear to have started shortly after we updated it to 1.6.1.

Kevin

On Thu, 29 Nov 2012, Derrick Brashear wrote:

> Of interest is that we're in osi_Sleep in GetDCache, which presumably
> means your cache is full.
>
>     2a103b6bc80 SLEEP    CV                      2
>                 afs_osi_Sleep+0x80
>                 afs_GetDCache+0x1724
>                 BPrefetch+0x68
>                 afs_BackgroundDaemon+0x2f4
>                 afsd_thread+0x1e8
>                 thread_start+4
>
>     3007524d7e0 SLEEP    CV                      2
>                 afs_osi_Sleep+0x80
>                 afs_GetDCache+0x1724
>                 afs_GetOnePage+0x490
>                 afs_getpage+0xe8
>                 fop_getpage+0x44
>                 segmap_fault+0x230
>                 afs_nfsrdwr+0x1130
>                 afs_vmread+0x70
>                 fop_read+0x20
>                 read+0x274
>                 syscall_trap+0xac
>
> So... something *should* be truncating the cache...
> but you don't appear to have a afs_CacheTruncateDaemon.
>
>
> Derrick
>