[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Sks-devel] Issues with sks recon - PTree database corrupt
From: |
Phil Pennock |
Subject: |
Re: [Sks-devel] Issues with sks recon - PTree database corrupt |
Date: |
Fri, 8 Jan 2010 13:34:18 -0800 |
On 2010-01-08 at 07:24 -0500, Mark Crichton wrote:
> Thanks to everyone who has offered to peer with me. I am now,
> unfortunately, having issues doing so.
>
> I've been looking at my logs, and it seems the "sks recon" process is
> dying on me with the last few lines in the log being:
>
> 2010-01-08 11:41:01 sks_recon, SKS version 1.1.1
> 2010-01-08 11:41:01 Copyright Yaron Minsky 2002-2003
> 2010-01-08 11:41:01 Licensed under GPL. See COPYING file for details
> 2010-01-08 11:41:01 Opening PTree database
> 2010-01-08 11:41:01 Setting up PTree data structure
> 2010-01-08 11:41:01 PTree setup complete
> 2010-01-08 11:41:01 Initiating catchup
> 2010-01-08 11:41:05 Raising Sys.Break -- PTree may be corrupted:
> Failure("remove_from_node: attempt to delete non-existant element from
> prefix tree")
> 2010-01-08 11:41:05 DB closed
>
> I am running sks-1.1.1 (using the source from Debian) and using DB 4.7.
> I saw on the list to try using db4.7_recover, which didn't work. I even
> shut down the sks server completely and rebuilt the ptree database. I
> still get the same error.
>
> Has anyone else seen this behavior, and know of a workaround?
Not seen it, but ...
Does the run-time user of sks have filesystem permission to access the
directory with the keydump in it, and all the parent directories?
The fast build requires that the original dump directory continues to
exist, and be accessible. There's a full build option which splits you
off from the dump.
-Phil
pgp8kRRbU6ffc.pgp
Description: PGP signature