[Netarchivesuite-users] wrong checksum

netarkivet netarkivet at statsbiblioteket.dk
Mon May 26 21:16:20 CEST 2008

Jobs that did not report right back to the system can be forced again to do so by copying the entire jobs-directory back to a harvester-directory and the restarting that specific harvester. 

This should have the job report back again to the system (and database).

It's a minor hack - and if the harvester previously completed the job so far as to generate the metadata.arc file this method has a minor issue because in that case it will generate a new metadata.arc file and try to upload that as well - that upload will fail because the harvester now tries to upload another file with another checksum. Just delete the second metadata.arc file from the job directory again once its moved back to /oldjobs

Bjarne Andersen

----- Original Message -----
From: Kåre Fiedler Christiansen <kfc at statsbiblioteket.dk>
Date: Monday, May 26, 2008 8:20 pm
Subject: Re: [Netarchivesuite-users] wrong checksum

> On Mon, 2008-05-26 at 19:41 +0200, Peter Moser wrote:
> > Hello!
> > 
> > I've got following warning in the systemmonitor. How critical is 
> it? Are 
> > the stored arc-Files consistant?
> It does sound serious. This message should only show up if the files
> have changed the checksum after the first initial upload! Something
> seems to have tampered with the files on your disk.
> How does your installation look? Do you have a replicated archive? In
> that case you should be able to recover from this situation using the
> bit preservation tools. Otherwise you may have to try to manually 
> figureout if your arc files are okay. Do you have a backup you can 
> comparewith?
> > Obviously the system is missing some files. Did heritrix 
> terminate 
> > correctly or not. In the job monitor the job is still "started", 
> > although no heritrix instance is running anymore.
> > What can I do?
> I think in this case the reason you have this problem is the lost
> connection you reported earlier. The database update that should have
> marked the job 'done' or 'failed' failed to update the database.
> Currently, unfortunately you cannot recover from this situation from
> within the system. You can shut down the system and manually update 
> thestattus field of the 'jobs' table in the database. or you can 
> ignore the
> seemingly started job - it's harmless.
> However, this should not cause the warning you see below. That warning
> seems to indicate something has changed on disk.
> The bit preservation interface should show you more.
> > Thanks in advance for your hints!
> I hope this covers your questions, otherwise I will gladly try to
> elaborate.
> Best,
>  Kåre
> _______________________________________________
> NetarchiveSuite-users mailing list
> NetarchiveSuite-users at lists.gforge.statsbiblioteket.dk
> https://lists.gforge.statsbiblioteket.dk/mailman/listinfo/netarchivesuite-users

More information about the NetarchiveSuite-users mailing list