Bug #46246
openmissing dep rhel 8: module perl-YAML:1.24:8010020190322131811:a5949e2e-0.x86_64 2
0%
Description
Problem: conflicting requests
2020-06-26T15:19:14.324 INFO:teuthology.orchestra.run.smithi087.stderr: - nothing provides module(perl:5.26) needed by module perl-YAML:1.24:8010020190322131811:a5949e2e-0.x86_64
2
Updated by Sebastian Wagner almost 4 years ago
- Assignee set to Ken Dreyer
No idea where this comes from. Note that this is Perl-YAML, which is afaik not used by cephadm nor ceph-ansible nor teuthology.
Ken, by chance do you have an idea where this comes from? Or do you know someone who could know?
Updated by Deepika Upadhyay almost 4 years ago
Sebastian Wagner wrote:
No idea where this comes from. Note that this is Perl-YAML, which is afaik not used by cephadm nor ceph-ansible nor teuthology.
Ken, by chance do you have an idea where this comes from? Or do you know someone who could know?
If not seen in recent teuthology runs, then maybe related to branch I am testing, will confirm.
Updated by Ken Dreyer almost 4 years ago
I thought some teuthology tests used Perl a long time ago (maybe for RGW?). I'm not sure if they still do.
Updated by Nathan Cutler almost 4 years ago
Ken Dreyer wrote:
I thought some teuthology tests used Perl a long time ago (maybe for RGW?). I'm not sure if they still do.
Yes, but apparently not this specific module (perl-YAML). At least, I could find no mention of it anywhere.
Updated by Sebastian Wagner over 3 years ago
- Status changed from New to Need More Info
Updated by Nathan Cutler over 3 years ago
- Project changed from Orchestrator to Ceph
- Status changed from Need More Info to New
- Assignee deleted (
Ken Dreyer) - Tags deleted (
low-hanging-fruit) - Severity changed from 3 - minor to 4 - irritation
IIRC the node provisioning process - the code of which resides at https://github.com/ceph/ceph-cm-ansible/ - installs Perl code directly from CPAN, and I suspect perl-YAML is getting pulled in as a dependency of CPAN.
I looked at the teuthology log. The job in question is:
tasks:
- install: null
- exec:
client.0:
- mkdir $TESTDIR/archive/ostest && cd $TESTDIR/archive/ostest && ulimit -Sn 16384
&& CEPH_ARGS="--no-log-to-stderr --log-file $TESTDIR/archive/ceph_test_objectstore.log
--debug-filestore 20 --debug-bluestore 20" ceph_test_objectstore --gtest_filter=-*/3
--gtest_catch_exceptions=0
- rm -rf $TESTDIR/archive/ostest
Which means the test does not even deploy a ceph cluster, much less use an orchestrator.
The job failed because ceph_test_objectstore failed. The RPM conflict with Perl-YAML came after the failure, during the "unwind" phase of the install task. Moving out of the Orchestrator project and lowering severity.