[blag-whereto] BLAG Trac #14: Sign packages in Slated then rebuild the repo

[BLAG]

#14: Sign packages in Slated then rebuild the repo
---------------------------------+------------------------------------------
 Reporter:  Homer                |       Owner:  Homer
     Type:  task                 |      Status:  new  
 Priority:  major                |   Milestone:       
Component:  BLAG Infrastructure  |     Version:       
 Keywords:                       |  
---------------------------------+------------------------------------------
 Packages hosted on Slated should be signed, and once this is done the repo
 will need to be rebuilt, since the yum indexes will then be out of sync.

 Step 1. Signing the packages.[[BR]]
 Status: This step is now complete.[[BR]]

 Step 2. Rebuilding the repo.[[BR]]
 Status: This step requires a local copy of a '''recent version''' of
 createrepo.[[BR]]

 Since this shared host is based on Centos 4.7, the most recent version of
 createrepo (and all its deps) available for Centos 4.7 is rather out of
 date, and does not (for example) utilise SQLite databases. This means I
 will need to rebuild all the various components from sources (alas,
 something I think I'll be doing on a permanent basis, until I can move to
 a dedicated host).

 An earlier attempt to do this failed, but this may have been because I
 lacked a recent copy of Python. Since the upgrade to Trac 0.11.3+ needed
 Python 2.5+, it is now installed, and I can therefore capitalise on this
 to try again.

 Earlier versions of createrepo had very little dependencies, but versions
 which support SQLite databases now have a vast dep list, including
 Mozilla's NSS and various cryptographic functions. The process of
 shoehorning rpm, yum and createrepo into this shared host will be slow and
 painful, and ultimately may not succeed at all, but I intend to make a
 best-effort attempt to do so anyway.

 The Trac system, and all other Slated systems, will be completely
 unaffected by this process, regardless of whether is succeeds or fails.

 Assuming the process ''does'' fail, then I will have no alternative but to
 rsync the entire tree to my local server, run the createrepo task from
 there, then resync the results back to Slated, but I'd '''really''' prefer
 not to do that, since my (home) bandwidth allowance is rather small, and I
 risk being subject to remedial measures by my ISP. For many reasons, it'd
 just be better if the whole workload was done on Slated, until such times
 as another server becomes available (FSF?).

 Will keep you updated with my progress.

-- 
Ticket URL: <http://trac.slated.org/ticket/14>
BLAG <http://www.blagblagblag.org/>
BLAG Linux and GNU