Re: [Sks-devel] keyserver.rainydayz.org back up

[John Clizbe]

Andy Ruddock wrote:
> Kristian Fiskerstrand wrote:
>> On 02/19/2013 06:19 PM, Andy Ruddock wrote:
>>> mid-December, so tried copying that but ran into "out of space
>>> for mutex" errors each time I tried to start sks - even after
>>> running "sks cleandb".

sks cleandb doesn't address this, it only affects the keys.
The problem here is in the actual database itself.

>> Did you try adding a DB_CONFIG file (as of 1.1.4 SKS allows this to
>> be used for new builds as well by having the file in the root DB
>> dir). Once this file is added you'll have to remove the BDB
>> environment (SKS will re-create it), see e.g [0]. That usually
>> solves mutex errors.

SKS 1.1.4 uses different default values for pagesize and ptree_pagesize
to try and alleviate the mutex exhaustion during db creation and loading.
A DB_CONFIG will partly help with this but the non-version specific option is
setting and increasing the pagesize parameters in sksconf from the defaults.

> There's a DB_CONFIG in the DB directory and another in the PTree
> directory. I don't remember at which point I copied them in - I
> possibly didn't clean the environment afterwards.

1.1.4 will use a DB_CONFIG at the time of database creation. For prior SKS
versions, one needs to remove the environment with db_recover after installing
DB_CONFIG in [K}DB and PTree. The environment should also be removed and
allowed to be recreated after changing DB_CONFIG.

>>> Anyway, back up and running - I hope the "sks build" bug gets 
>>> fixed, It's been reported and exists within their bug reporting 
>>> system.

I committed changes for both parts of this in SKS 1.1.4

>> For what it is worth I don't see anything in the SKS BTS at [1].
>> Which version of BDB you're using would be interesting to know to
>> try to replicate (in SKS 1.1.4 try `sks version` to get this
>> information)
> 
> I'm using the version that ships with Debian Wheezy (testing), which
> at the moment has the version number libdb5.1.29-5
> I'm not sure how that relates to upstream and what patches may have
> been applied.
> 
> The issue is in the Debian bug tracking system :
> 
> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=699848
> 
> I presume they would have reported any bugs back upstream.

I only find out about SKS bug reports against Debian when someone mentions
them on this list. I replied to that bug and suggested a workaround to
the OP. It wasn't until I read that bug that it clicked that this was most
likely also your problem.

> Considering that the machines I use as "servers" are Pentium P4
> machines, that between them host mail, web, ftp, ssh & sks services -
> I'm impressed that they handled the workload as well as they did.
> 
> (I don't expect either machine to last very much longer and am in the
> process of investigating the best way to replace them)

My Windows SKS instance runs on a P-III ThinkPad ;-)

-John

-- 
John P. Clizbe                      Inet: John (a) Gingerbear DAWT net
SKS/Enigmail/PGP-EKP                  or: John ( @ ) Enigmail DAWT net
FSF Assoc #995 / FSFE Fellow #1797  hkp://keyserver.gingerbear.net  or
     mailto:address@hidden

Q:"Just how do the residents of Haiku, Hawai'i hold conversations?"
A:"An odd melody / island voices on the winds / surplus of vowels"

signature.asc
Description: OpenPGP digital signature