https://www.gridpp.ac.uk/w/index.php?title=SE_Lost_Disk-Server&feed=atom&action=historySE Lost Disk-Server - Revision history2024-03-29T11:34:38ZRevision history for this page on the wikiMediaWiki 1.22.0https://www.gridpp.ac.uk/w/index.php?title=SE_Lost_Disk-Server&diff=2346&oldid=prevChristopher walker at 16:06, 14 April 20122012-04-14T16:06:50Z<p></p>
<p><b>New page</b></p><div>== SE Lost Disk Server Procedure ==<br />
<br />
The following instructions are valid for ''any'' Storage Element, and form recommended best practice for communication with the VO and removing file remanants from your SRM. These instructions are explicitly for when files are permanently lsot and are not recoverable by the site admin. Instructions for temporarily unavailable files/disk-servers are different.<br />
<br />
== Overview ==<br />
* List all files on the diskserver /filesystem lost.<br />
* Prepare list of SURLs for the VOs in question.<br />
* Supply to VO via appropiate route this list of SURLs statiting that theirs are lost.<br />
* Remove entries from databases on the SRM.<br />
*Inform VO that all old entries have been removed and that they can either replicate the files from another site or update their catlaogs.<br />
<br />
== Listing files on diskserver and preparing list of SURLs for the VO ==<br />
*CASTOR<br />
*DPM<br />
*dCache<br />
<br />
=== StoRM ===<br />
==== Lustre ====<br />
<br />
See: https://www.gridpp.ac.uk/wiki/StoRM#Generating_a_list_of_SURLS<br />
<br />
==== GPFS ====<br />
<br />
== VO specific communication routes ==<br />
In general; VOs can be contacted by submission of a GGUS ticket which will be assigned to the VOSupport queue and assigned to the VO in question. Some of our major VOs have extra steps which can be used as an alternative communication method(s) if you wish to speed up the process.<br />
*ATLAS<br />
*CMS<br />
*LHCB<br />
== Removing entries from the database ==<br />
*CASTOR<br />
**Additional issues with castor are that there may well be other repliacs on tape or disk of the file may exist. These should have been checked to seee if the file is recoverable ( via internal site commands.). If not then the databases should be cleaned using stager_rm and nsrm appropiately.<br />
*DPM<br />
*dCache<br />
=== StoRM ===<br />
If files are not on the filesystem, they are not available via StoRM. Note that file access to a storage server that is down will block in Lustre (unless Lustre is configured so that they fail).</div>Christopher walker