Because 2G for /var is, quoting @faidon, 'bonkers'.
How about a unified min 30G /, and rest of space available via LVM? That should be enough, and we have enough disk space too.
This is an unnecessary deviation from prod (which has big enough /var, usually). Causes a lot of configuration headaches when none should be required.
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Open | None | T53494 Use Beta cluster as a true canary for code deployments (epic) | |||
| Open | None | T87220 Minimize infrastructure differences between Beta Cluster and production | |||
| Resolved | yuvipanda | T87214 Migrate beta cluster jobrunner to HAT | |||
| Resolved | mmodell | T87003 New disk partition scheme for labs instances | |||
| Resolved | thcipriani | T88304 Setup staging-palladium as puppetmaster and saltmaster | |||
| Resolved | faidon | T90229 Package geoipupdate for jessie |
Let's do this for the Jessie images to start with, and then make precise / trusty images like this too.
We should do this for Trusty and Precise too. I've to re-image deployment-jobrunner01 to be HAT...
Change 186113 had a related patch set uploaded (by Yuvipanda):
Scrap the /var/log lvm partition; make the / 20G by default.
Change 186113 had a related patch set uploaded (by Andrew Bogott):
Scrap the /var/log lvm partition; make the / 20G by default.
Change 186113 merged by Andrew Bogott:
Scrap the /var/log lvm partition; make the / 20G by default.
Change 186121 had a related patch set uploaded (by Yuvipanda):
labs_vmbuilder: Remove separate /var and /var/log
Change 186121 merged by Andrew Bogott:
labs_vmbuilder: Remove separate /var* and increase / to 19G
How is this resolved? Did Andrew build the new images and make them
available via glance?
So, precise and trusty images seem to work! I've deprecated the old images and made new ones. And there was much rejoicing!