Ceph QA

This looks like a test issue to me. No tests were changed as part of this PR.

One pool was created and destroyed during the test:

teuthology.log|2025-07-26T08:59:16.374 DEBUG:teuthology.orchestra.run.smithi001:> sudo adjust-ulimits ceph-coverage /home/ubuntu/cephtest/archive/coverage timeout 120 ceph --cluster ceph osd pool create unique_pool_0 16 16 erasure teuthologyprofile
teuthology.log|2025-07-26T08:59:18.392 DEBUG:teuthology.orchestra.run.smithi001:> sudo adjust-ulimits ceph-coverage /home/ubuntu/cephtest/archive/coverage timeout 120 ceph --cluster ceph osd pool application enable unique_pool_0 rados --yes-i-really-mean-it || true
teuthology.log|2025-07-26T09:09:22.930 DEBUG:teuthology.orchestra.run.smithi001:> sudo adjust-ulimits ceph-coverage /home/ubuntu/cephtest/archive/coverage timeout 120 ceph --cluster ceph osd pool rm unique_pool_0 unique_pool_0 --yes-i-really-really-mean-it

The thrasher, however, still seems to be running and gets upset about there being no PGs:

teuthology.log|2025-07-26T09:09:22.593 INFO:tasks.thrashosds.thrasher:pool unique_pool_0 min_size is 2
teuthology.log|2025-07-26T09:09:23.284 INFO:tasks.thrashosds.thrasher:pool unique_pool_0 local_k=2 local_m=1
teuthology.log|2025-07-26T09:09:23.284 INFO:tasks.thrashosds.thrasher:setting k=2 from previous 0
teuthology.log|2025-07-26T09:09:23.284 INFO:tasks.thrashosds.thrasher:setting m=1 from previous 99
teuthology.log|2025-07-26T09:09:23.284 INFO:tasks.thrashosds.thrasher:using k=2, m=1
teuthology.log|2025-07-26T09:09:23.284 INFO:tasks.thrashosds.thrasher:dead_osds=[], live_osds=[0, 8, 12, 5, 9, 13, 2, 6, 10, 3, 11, 15, 4, 7, 14, 1]
teuthology.log|2025-07-26T09:09:23.284 INFO:tasks.thrashosds.thrasher:chose to kill 1 OSDs
teuthology.log|2025-07-26T09:09:23.627 INFO:tasks.thrashosds.thrasher:No pgs; trying again
teuthology.log|2025-07-26T09:09:28.990 INFO:tasks.thrashosds.thrasher:No pgs; trying again
teuthology.log|2025-07-26T09:09:34.349 INFO:tasks.thrashosds.thrasher:No pgs; trying again
teuthology.log|2025-07-26T09:09:34.351 INFO:tasks.thrashosds.thrasher:Traceback (most recent call last):

So I think this is an unfortunate timing window:

1. Thrasher detects a pool exists. 
2. Pool is deleted
3. Thrasher attempts to detect PGs in the now deleted pool, which it still remembers.