I was seeing this last week after updating the CA rpms, I found some
posting to dcache-user-forum suggesting that it could be a result of
"the too many CA's bug" though I thought that was only client side and
only affected gsidcap transfers.
However, I went ahead and removed the ca_SWITCH* rpms and I've only seen
this happen once since then (and that was the day after).
Your milage may vary!
Chris.
> -----Original Message-----
> From: GRIDPP2: Deployment and support of SRM and local
> storage management [mailto:[log in to unmask]] On
> Behalf Of Matt Doidge
> Sent: 07 July 2006 10:38
> To: [log in to unmask]
> Subject: Lancaster's new dcache woe-end of file error
>
> Hello all,
> I suppose it's all been going smoothly for too long here at
> Lancs. Our dcache SE has started going flakey on us, we find
> that attempts to transfer files, using lcg-cp or even plain
> srmcp give "end of file"
> error messages and deliver empty files. A restart of the
> dcache services fixed this the first time it happened on
> Wednesday. Then we keeled over again later that night. So I
> made sure everything was up to date, and restarted it all
> again. We made sure transfers worked, and left it. Then it
> broke again this morning. I'm not sure what's wrong.
>
> Our symptoms are similar to the ones detailed on this thread
> on the dcache site:
> http://www.dcache.org/archive/user-forum/0307.shtml
>
> However, we have no OutofMemory type errors in our logs as
> was suggested to look for in that thread.
>
> What we do see in our dCacheDomain.log is many messages of
> the flavour:
> 07/07 09:27:34 Cell(PoolManager@dCacheDomain) : PnfsHandler :
> CacheException (4) : Pnfs error : Pnfs lookup failed
>
> Any ideas? I'm stuck.
> Thanks in advance,
> Matt
>
|