[ZCM] [ZC] 1420/ 5 Comment "Corrupted Data.fs"

[Collector: Zope Bugs, Features, and Patches ...]

Issue #1420 Update (Comment) "Corrupted Data.fs"
 Status Accepted, Database/bug medium
To followup, visit:
  http://collector.zope.org/Zope/1420

==============================================================
= Comment - Entry #5 by h.rauch on Jul 23, 2004 2:02 am

Sometime good news are bad news!

After hours of investigations, we finally found the source for the dameged Data.fs. It was a faulty sata-raid-controller. It worked well under normal conditions, but wasn't able to do the same under heavy load. Maybe the member chips is faulty - I don't know. 

Thanks for your help!

Hans

________________________________________
= Comment - Entry #4 by tim_one on Jul 15, 2004 12:02 pm

Hans, reports of corruption are very rare, and no general cause is known.  Please read the info here first, and see whether it helps you:

http://zope.org/Wikis/ZODB/FileStorageBackup

Note that packing is in no way an error recovery or error prevention procedure (its purpose is to save space).
________________________________________
= Accept - Entry #3 by tim_one on Jul 15, 2004 11:05 am

 Status: Rejected => Accepted

 Supporters added: tim_one

Please don't close FileStorage corruption reports.  We've explicitly asked people to report them here, and they're so complicated it's impossible to keep them straight on a mailing list.

Sending the info to a mailing list too is a good idea, for community feedback, but only a Collector item can record progress and clues in a usable way over time.
________________________________________
= Reject - Entry #2 by ajung on Jul 15, 2004 10:52 am

 Status: Pending => Rejected

Your report is a problem report but not a bug report.
Also your issue does not contain much useful informations
to help you. I suggest to bring this issue up on the zodb-dev mailinglist where a large audience might help you.
________________________________________
= Request - Entry #1 by Anonymous User on Jul 15, 2004 10:21 am

Two weeks ago I installed a new server: Linux SuSE 9.1 - 64-Bit version, SMP 2 x Opteron, 2GB, Sata-Raid-System. There doen't seem to be any hardware failure, since the other servers (MySQL, Aapache ..) seem to work fine.

I'm using a ZEO-Server, five Zope-Clients with different ports and an apache-server infront of these Zope clients. This configuration worked fine with the old system: 4 Xeon CPUs, 2GB, SCSI-Raid. Size of Data.fs: 330 MB.

Every 10, 20, 30 hours I have problems, when I try to pack Data.fs. I got unpickle-errors and lots of other strange errors. When I remove these objects, I'm able to pack the database, but the errors will come back again. These corrupted objects are mostly old items.

I'm now starting the packing process every hour for minimal data loss. But this does not solve my problem.

The url is http://portal.bildung.hessen.de



==============================================================