Re: [Errno 256] No more mirrors to try

> May I ask if you have more news on a possible fix? I would just like to know if we can expect a fix soon.
> People installing gLite in SL4 rely on jpackage and this is now a big showstopper for sys admins abd developers
> to automate installations. If a fix is eventually available, what is the time scale to see it deployed in the repo?


Hi,

though somebody has already described the current condition in the
LCG-ROLLOUT list, I believe it's worthy to repeat it here.

In fact (as far as I know), the bug has already been fixed. But
another problem appeared.

# yum update

=============================================================================
 Package                 Arch       Version          Repository        Size
=============================================================================
Updating:
 curl                    i386       7.12.1-11.2.el4_8.3  sl-errata         231 k
 curl-devel              i386       7.12.1-11.2.el4_8.3  sl-errata         246 k
(...)

Transaction Summary
=============================================================================
Install      0 Package(s)
Update       7 Package(s)
Remove       0 Package(s)
Total download size: 5.5 M
Is this ok [y/N]:

Looks familiar. But let me answer "N" and do "yum clean all" instead.

# yum clean all
Cleaning up Everything
31 headers removed
(...)

The next "yum update" results in:

# yum update

  File "repos.py", line 287, in populateSack
  File "sqlitecache.py", line 96, in getPrimary
  File "sqlitecache.py", line 89, in _getbase
  File "sqlitecache.py", line 359, in updateSqliteCache
  File "sqlitecache.py", line 251, in addPrimary
  File "sqlitecache.py", line 197, in insertHash
  File "sqlitecache.py", line 449, in values
  File "sqlitecache.py", line 441, in __getitem__
  File "/opt/glite/lib/python2.3/site-packages/amga/mdparser.py", line
73, in __getitem__
    # Check whether it is in our parser dictionary
KeyError: 'sourcerpm'

So those unfortunate folks (including me) who did a "yum clean all"
inbetween (as was suggested on the LCG-ROLLOUT list a few days ago)
are now out of luck.

Peter Slizik