[JIRA] (OVIRT-921) Smarter cache handling for mock_runner.sh
Barak Korren (oVirt JIRA)
jira at ovirt-jira.atlassian.net
Thu Dec 8 09:20:02 UTC 2016
[ https://ovirt-jira.atlassian.net/browse/OVIRT-921?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Barak Korren updated OVIRT-921:
-------------------------------
Labels: mock mock_runner.sh performance standard-ci (was: )
> Smarter cache handling for mock_runner.sh
> -----------------------------------------
>
> Key: OVIRT-921
> URL: https://ovirt-jira.atlassian.net/browse/OVIRT-921
> Project: oVirt - virtualization made easy
> Issue Type: Improvement
> Components: Jenkins
> Reporter: Barak Korren
> Assignee: infra
> Labels: mock, mock_runner.sh, performance, standard-ci
>
> We've improved our handing of mock caches in OVIRT-902, but it could be better.
> Right now we're just deleting everything from {{mock_cleanup.sh}} every two days, but we could do things in a smarter and more refined way:
> * Mock actually has two separate caches that could be handled separately:
> *# The yum cache - Caches package that are installed in the chroot
> *# The root cache - Keeps an archive of the whole chroot
> * We can make the yum cache shared by all the chroots that use the same repos
> * We can keep the yum cache for much longer since 'yum' will refresh it automatically. We can tune this from mock root config file:
> ** {{config_opts\['plugin_conf']\['yum_cache_opts']\['max_age_days']}}
> ** {{config_opts\['plugin_conf']\['yum_cache_opts']\['max_metadata_age_days']}}
> * We can also control the chroot cache from the mock config file, this will allow us greated flexibility, like deciding that some chroots live longer:
> ** {{config_opts\['plugin_conf']\['root_cache_opts']\['max_age_days']}}
> * We can use a different compression algorithm for the chroot acche, this may yield better performance.
> * We still need some cleanup in {{moch_cleanup.sh}} but only for chroot caches that ahvn't been in use for a long time.
--
This message was sent by Atlassian JIRA
(v1000.620.0#100023)
More information about the Infra
mailing list