Sepia January 2020

sepia@ceph.io

10 participants
4 discussions

by David Galloway

Hi all, I'm guessing this question is mostly for Sage but maybe others have input. As you should be aware, we are quickly running out of space in the Sepia lab. I have room for the incoming Ceph Foundation hardware but anything beyond that is going to require purchasing new racks (which will take months) or retiring old equipment. Currently, I need to make room for the 6 systems purchased to replace the rex systems that were in the LA office. I'm wondering if there would be any objection to retiring the apama systems. Only one of them is actually doing anything useful -- it has some OSDs in the LRC. It only has 48GB RAM and the 7 OSDs currently running on it take up all of it. https://wiki.sepia.ceph.com/doku.php?id=hardware:apama "But David, why not just get rid of 3 mira instead?" The miras are in racks that only have 1Gb networking. I'd need to throw out 4 *and* buy a 10Gb switch. The apama are in the "infra" rack which already has a 10Gb switch. Thoughts? -- David Galloway Systems Administrator, RDU Ceph Engineering IRC: dgalloway

4 years, 2 months

got ScheduleFailError while trigger teuthololgy tests

by Rishabh Dave

Hi all, For two of my builds[1][2] I get ScheduleFailError while trying to trigger the tests. Following is the error message I receive on executing the teuthology-suite command to trigger tests - teuthology.exceptions.ScheduleFailError: Scheduling rishabh-2020-01-17_16:21:33-fs-wip-rishabh-fs-qa-UnicodeDecodeError-distro-basic-smithi failed: Failed to fetch package version from https://shaman.ceph.com/api/search/?status=ready&project=ceph&flavor=defaul… I've copied the entire output here[3]. It includes the traceback and few more lines in addition to the error message above. Is anyone else experiencing this or is it just me? Is there a known workaround? I am planning to retrigger builds otherwise since I can't trigger any tests using these builds. Thanks, - Rishabh [1] https://shaman.ceph.com/repos/ceph/wip-rishabh-xfstests-dev-on-centos8/0755… [2] https://shaman.ceph.com/repos/ceph/wip-rishabh-fs-qa-UnicodeDecodeError/e0a… [3] https://gist.github.com/rishabh-d-dave/7d3481fabb6d2c3f413906fa41669983

4 years, 3 months

all mira are locked...

by Sage Weil

mira 1 0 dgallowa@dgallowa mira 1 1 sseshasa@teuthology mira 1 1 vasu@teuthology mira 1 1 zack(a)zwork.local mira 1 1 scheduled_zack@teuthology mira 2 2 smanjara@teuthology mira 2 0 scheduled_kchai@teuthology mira 3 3 yehudasa@yehudasa-desktop mira 3 3 dis@zambezi mira 3 1 jenkins-build@teuthology mira 3 3 joshd@teuthology mira 3 3 bhubbard@teuthology mira 4 3 zack@teuthology mira 4 4 gmeno(a)redhat.com mira 4 4 VPSHOST@VPSHOST mira 6 4 dgalloway@w541 mira 10 0 (free) mira 14 13 LONG_RUNNING_CLUSTER mira 26 4 jenkins@jenkins mira 33 33 vshankar Anybody not using what they have locked that can unlock? (vshankar, are you using all 33 of those?) thanks! sage

4 years, 3 months

issue with triggering teuthology tests for centos8

by Rishabh Dave

Hello, I am trying to test this PR[1] and have builds ready for that[2] but on running the following command I get this weird issue[3] - teuthology-suite -v -m smithi -c xfstests-dev-on-centos8 -s kcephfs -p 50 --filter acls-kernel-client Nothing is wrong with the build or the branch per se; I confirmed that by triggering a different test suite[4] which was successful. From the tracker ticket[5] for the PR it looks like this testsuite was ran before and it was triggered successful then. From the stdout[3] it feels like the issue lies in teuthology code and triggering a job for CentOS 8 triggers it. So I'm looking at the teuthology code right now, but if this has occurred before and someone knows the cause, fix or someone else who is fixing this already, please let me know. Thanks, - Rishabh [1] https://github.com/ceph/ceph/pull/32524 [2] https://shaman.ceph.com/builds/ceph/xfstests-dev-on-centos8/ [3] https://gist.github.com/rishabh-d-dave/265c0e044c50688bd4da18b28540d7f0 [4] http://pulpito.front.sepia.ceph.com/rishabh-2020-01-07_17:29:02-fs-xfstests… [5] https://tracker.ceph.com/issues/43486

4 years, 3 months

2024

2023

2022

2021

2020

2019

Sepia January 2020