Hi Andy,
What you say is true, but the client should still have gone back to the
redirector to ask for a refresh, no? (i.e. removing like this induces
an extra redirect, but the collaboration of the client in the system should
mean that things are self-correcting.)
Pete
On Thu, Apr 07, 2005 at 02:25:53AM -0700, Andrew Hanushevsky wrote:
> Hi Derek,
>
> >From the description, you did an out-of-band delete, yes? That is you
> simply rm'd the file by hand. If so, the system knew nothing about the
> deletion. If you used xrootd to delete the file, it would have notified
> the olbd that the file was gone and the situation would not have occured.
> So, was the file manually removed outside the context of the system?
>
> Andy
>
> On Thu, 7 Apr 2005, Derek Feichtinger wrote:
>
> > Hi,
> >
> > I tried the following:
> >
> > setup:
> > one redirector, 2 ro leaf nodes, same file "fileA"on the leaf nodes.
> >
> > 1. Reading fileA via redirector nicely redirects me alternatingly to both
> > leaf nodes
> > 2. I remove fileA from leaf-node-1
> > 3. Reading fileA via redirector keeps sending me every other time to the
> > machine which no longer has the file and I get "Error accessing path/file".
> >
> > I understand that the client should handle the notification (cache resync) to
> > the olbd master?
> >
> > Cheers,
> > Derek
> >
> > --
> > Dr. Derek Feichtinger Tel: +41 22 767 10 07
> > LCG/ARDA Group email: [log in to unmask]
> > CERN http://people.web.psi.ch/feichtinger
> > CH-1211 Genève 23
> >
> >
> >
-------------------------------------------------------------------------
Peter Elmer E-mail: [log in to unmask] Phone: +41 (22) 767-4644
Address: CERN Division PPE, Bat. 32 2C-14, CH-1211 Geneva 23, Switzerland
-------------------------------------------------------------------------
