Re: [Zope-dev] CorruptedDataError sniff !

2005-06-02 Thread Tim Peters
[Eric Brun]
| Houston, we've got big a problem !

 Here is the traceback in ZEO_EVENTS.log:
 ...
 CorruptedDataError: 968647571
 

 And I can't pack the zodb, ...

Packing is the last thing you should be trying when you have
corruption.  There's no mechanical, safe way to recover.  Please read
this:

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

and start by using fstest, then fsrefs, as described there.

 I run  Zope 2.6.2 , ZEO 2 and ZODB 3.1

Just noting that it's harder to get help with old releases.

 I have tried to do a fsrecover, it removes a transaction. Then when I tried 
 to lunch
 a pack,

Packing isn't an error recovery procedure.

 I've got a very strange error about unpickling :

That suggests object pickles are corrupt too; fstest can't discover
that, and fsrecover can't repair that.  fsrefs _may_ (or may not) be
provoked into complaining, or even crashing, by it.

 Error unpickling, ((UBTrees._IIBTreeqU... [lots of gibberish] ...

 I try to copyTransaction in a new Storage but it failed too.

This is a surprise wink?  Seriously, garbage in, garbage out
applies here.  The garbage has to be repaired first.

 This is a production site.
 My last backup which is not corrupted date of Sunday morning !

Since fsrecover didn't work for you, you need to learn more about the
nature of the damage.  fstest, fsrefs, and fsdump are the basic tools
to start figuring that out.
___
Zope-Dev maillist  -  Zope-Dev@zope.org
http://mail.zope.org/mailman/listinfo/zope-dev
**  No cross posts or HTML encoding!  **
(Related lists -
 http://mail.zope.org/mailman/listinfo/zope-announce
 http://mail.zope.org/mailman/listinfo/zope )


Re: Re: [Zope-dev] CorruptedDataError sniff !

2005-06-02 Thread Eric Brun
 
 Packing is the last thing you should be trying when you have
 corruption. 

In fact I discover the corruption during the pack.

 There's no mechanical, safe way to recover.  Please read
 this:
 
 http://zope.org/Wikis/ZODB/FileStorageBackup

This is a great document. I think the origin of the corruption is my RAID 
controler so I have ask to the server owner to switch for a new server.

 
 and start by using fstest, then fsrefs, as described there.



 
  I run  Zope 2.6.2 , ZEO 2 and ZODB 3.1
 
 Just noting that it's harder to get help with old releases.

I have mount my zodb on a Zope2.7.6 and I'm working on this release to solve my 
problem.

fstest say all is ok.
fsrefs say that :

oid 0x37d8e5 BTrees._IIBTree.IIBTree
last updated: 2005-05-28 15:48:11.159195, tid=0x35D95942F9CD111L
refers to invalid object:
oid 0x5d4787 missing: 'BTrees._IIBTree.IIBucket'

oid 0x37d8ec BTrees._IOBTree.IOBucket
last updated: 2005-05-07 00:56:12.524043, tid=0x35D1BF8356F9844L
refers to invalid object:
oid 0x39c22b missing: 'BTrees._IIBTree.IIBTree'


..

etc



___
Zope-Dev maillist  -  Zope-Dev@zope.org
http://mail.zope.org/mailman/listinfo/zope-dev
**  No cross posts or HTML encoding!  **
(Related lists -
 http://mail.zope.org/mailman/listinfo/zope-announce
 http://mail.zope.org/mailman/listinfo/zope )