Closed HenyPotter closed 3 years ago
Seems we hit some kernel bug.
Something gets stuck in a syscall loop and runs at 100% on 1 CPU:
(logging in, no process shows any CPU usage more than a percent or so, so very freaky)
We also see at this point, we get a memory leak where our (un+)reclaimable memory both start heading up.
Feb 27 06:46:22 redis1 update-apt-xapi[8418]: segfault at 24 ip 000000000052a36c sp 00007fff26997860 error 6 in python2.7[400000+2bc000]
Feb 27 06:46:22 redis1 BUG: unable to handle kernel paging request at fffffff47824f0a0
Feb 27 06:46:22 redis1 IP: [<ffffffffa007c59f>] drm_mm_insert_node_in_range_generic+0x1af/0x340 [drm]
Feb 27 06:46:22 redis1 PGD 1c11067 PUD 0
Feb 27 06:46:22 redis1 Oops: 0000 [#2] SMP
It appears syslog was killed around this point (probably an OOM). First the Username -> UUID cache died, but the crash on the session data was more impactful.
We got some assistance with the ratelimits so things should return to normal "soon".
Okay, thank you and have a nice day! :)
Still does not work on some UUIDs or names, Notch is working but KuCZik not
Still seems to be a problem. Getting Steve heads for valid uuids.
https://minotar.net/helm/6f23cfa0b8e34c35855bdfb94b4f8e91/100.png
This should be reopened.
Hi @harvanchik , I've been checking this off and on the last few days and I haven't seen an issue loading the helm. Are you able to screenshot the error you are seeing?