Page MenuHomePhabricator

Request volume quota increase for integration Cloud VPS project
Closed, ResolvedPublic

Description

Project Name: integration

Type of quota increase requested: volume quota increase

Reason:

The integration-castor03 is based on Debian Stretch and has to be replaced with a Bullseye image (T252071). It uses a custom flavor to get extra disk space which is no more available via Horizon, we thus a new flavor to be created. That is also an opportunity to revisit its sizing.

We currently have:
g2.cores8.ram16.disk160

VCPUs8
RAM16G
Disk160G

Looking at the Grafana board for integration-castor03:

  • there is barely any CPU usage, so we can keep 8 vCPUs
  • RAM is filed by disk cache. I think we would benefit more RAM for that
  • the partition holding the cache is 138G and is full. Beside we are not too good at garbage collecting the cache, we also have more repositories and jobs and would need an extension
  • we could use a bump of IO throttling

We would need to create a new instance and offload the disk storage to a volume set to 320G (double the current size). The integration project currently uses 53G out of a 80G disk quota. May you increase the volume quota from 80G to 400G?

Event Timeline

For this specific use case we are probable better off creating a volume via Horizon and attach it to the instance. That gives us the liberty to resize however we need and saves WMCS people from having to create yet another custom flavor.

The integration has a volume quota though: Used 53GB of 80GB. May we get that one bumped to 400GB?

hashar renamed this task from Request new flavor for integration Cloud VPS project to Request volume quota increase for integration Cloud VPS project.Mar 28 2022, 9:48 AM
hashar updated the task description. (Show Details)
Andrew claimed this task.
Andrew added a subscriber: Andrew.

done!

As a followup, I had to recreate the instance and reattach the volume but it was not magically mounted to /srv. Andrew said prepare-cinder-volume creates an fstab entry but since I have recreated the instance it was gone.

I thus attached the volume to the instance and applied profile::labs::cindermount::srv.

Notice: /Stage[main]/Profile::Labs::Cindermount::Srv/Cinderutils::Ensure[cinder_on_srv]/Exec[prepare_cinder_volume_/srv]/returns: executed successfully
$ df -h  /srv
Filesystem      Size  Used Avail Use% Mounted on
/dev/sdb        314G  205G   94G  69% /srv

Magic.

Change 961844 had a related patch set uploaded (by Hashar; author: Hashar):

[operations/puppet@production] ci: manage cinder volume on Castor instance

https://gerrit.wikimedia.org/r/961844

Change 961844 merged by David Caro:

[operations/puppet@production] ci: manage cinder volume on Castor instance

https://gerrit.wikimedia.org/r/961844