Issue #21103 has been updated by Klavs Klavsen.
I just tested setting metadata_expire=1.. Problem is EVERY time yum is used, and puppet uses yum a LOT to query just about what's installed (yum list installed..) - but even then, yum checks updates of mirrors with metadata_expire=1.. Even though the information requested should mean no need for updating of metadata (since I'm asking for installed only).. So yes - YUM is broken! I'm curious as to how this is handled with APT - which is equally broken - except it's just well known - that one HAS to run apt-get update.. whereas yum automaticly does this by itself (normally every 6 hours according to man page). For now, I'm going with metadata_expire=1m - to let my puppet run not be too annoyed by this, and updates still be known within a minute (meaning I have to put in a sleep 60 in my release scripts that wants to push out packages). IMHO if Puppet wants to be sure that it sees the latest information (which would clearly be something it should ensure, atleast when it a yum install fails) - it should ENSURE to update metadata at the beginning of it's run - no matter wether it's YUM or APT we're talking - they both have the same issue - and unfortunately it's not fixable in the manifests, unless you want to expire cache on EVERY run AND live with the fact, that this would register as a change on the client on every run :( ---------------------------------------- Feature #21103: Add puppet yum provider retry operation https://projects.puppetlabs.com/issues/21103#change-98940 * Author: Bogdan Dobrelya * Status: In Topic Branch Pending Review * Priority: Normal * Assignee: Bogdan Dobrelya * Category: * Target version: * Affected Puppet version: * Keywords: yum provider * Branch: https://github.com/puppetlabs/puppet/pull/1691 ---------------------------------------- Without this feature, in case of "Error 256 No more mirrors to try" yum provider would have failed operation in progress which would cause, for example, wanted package installation failed and all dependencies skipped. This is a problem because yum should provide operation completion a better way, at least it should try to expire local cache and retry the operation once or several times more. Solution/Improvement: If operation wanted had failed, puppet provider for yum would try to expire cache and redo failed operation for several times (3 is default). Why it is a good improvement: That might restore failed operation in some cases (f.e. broken or outdated metadata file in cache) and allow puppet to continue operations. -- You have received this notification because you have either subscribed to it, or are involved in it. To change your notification preferences, please click here: http://projects.puppetlabs.com/my/account -- You received this message because you are subscribed to the Google Groups "Puppet Bugs" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/puppet-bugs. For more options, visit https://groups.google.com/groups/opt_out.
