Page MenuHomePhabricator

Investigate I/O and implement cgroups on stat hosts
Closed, ResolvedPublic

Description

Creating this ticket to:

Specific tasks for stat1011:

[] Use the Prometheus metrics recorded around the time of the last stat1011 hang to estimate the upper bounds of I/O performance on stat1011 in its current configuration
[] Get hardware information on
-- stat1011's host bus adapter ( the Dell HBA355i Front, product info here )
-- stat1011's disk . This information will help us compare theoretical maximums to what we saw during the last outage.

Event Timeline

bking changed the task status from Open to In Progress.Oct 7 2024, 8:22 PM
bking claimed this task.
bking triaged this task as Medium priority.
bking updated the task description. (Show Details)
bking updated the task description. (Show Details)

This writeup from Facebook provides an excellent real-world example of using cgroups v2 to protect workloads.

bking renamed this task from Investigate I/O and implement cgroups on stat1011 to Investigate I/O and implement cgroups on stat hosts.Oct 9 2024, 3:10 PM
bking updated the task description. (Show Details)

Change #1079281 had a related patch set uploaded (by Bking; author: Bking):

[operations/puppet@production] stat hosts: create/enable cgroups for memory and i/o

https://gerrit.wikimedia.org/r/1079281

Change #1079281 merged by Bking:

[operations/puppet@production] stat hosts: create/enable cgroups for memory and i/o

https://gerrit.wikimedia.org/r/1079281

Per the above PR, we have activated memory and I/O cgroups on all stat hosts. I've crossed out the rest of the AC as it's entirely possible that we'll be using Ceph homedirs instead of the current disks fairly soon (ref: this design doc). We can always take a closer look at the disks if necessary, but I'm going to close this one out for now.