Page MenuHomePhabricator

Request increased quota for text-to-speech Cloud VPS project
Closed, ResolvedPublic

Description

Project Name: text-to-speech

Type of quota increase requested: CPU, RAM

Amount to increase: Increase CPU from 8 to 18, increase RAM from 16GB to 36GB

Reason:

  • To upgrade larynx-be-01 from g3.cores8.ram16.disk20 to g3.cores16.ram32.disk20 as it currently sits at less than 1GB RAM free when idle — admittedly most RAM is "held" by the running Python processes, but would like to start performance testing the larynx text-to-speech engine
  • To create a g3.cores2.ram4.disk20 instance for monitoring

Event Timeline

+1, assuming there's a typo and the desired amounts are 16 rather than 18 for CPU, 32 rather than 36 for memory? Or is some overhead required?

Oh reading again, I think the overhead is for the additional instance, makes sense!

Mentioned in SAL (#wikimedia-cloud-feed) [2022-11-25T11:45:26Z] <wm-bot2> Increased quotas by 10 cores, 20 ram (T323617) - cookbook ran by arturo@nostromo

aborrero claimed this task.

Hi @aborrero — it looks like the CPU quota increased but not the RAM

image.png (268×784 px, 19 KB)

(and while I have you, does "resizing" an instance preserve the state of the VM or is it akin to a rebuild?)

Ok, investigating.

(and while I have you, does "resizing" an instance preserve the state of the VM or is it akin to a rebuild?)

Resize should work safely. Please open a ticket if it does not.

Mentioned in SAL (#wikimedia-cloud-feed) [2022-11-29T10:46:55Z] <wm-bot2> Increased quotas by 20480 ram (T323617) - cookbook ran by arturo@nostromo

Sorry, I was confused by the units. Should be better now!