For some reason the requests library has trouble connecting from
ubuntu 18.04. I reproduced this locally on my 18.04 desktop, although
there it fails on the first API request instead of the last (as in QA).
In any case, this appears to be a client library problem.
Fixes: https://tracker.ceph.com/issues/43720
Signed-off-by: Sage Weil <sage@redhat.com>
This was broken by the variable PG log lengths in
9c69c2f7cc585b5e13e4d1b0432016d38135a3de.
Disable the new option to get (roughly) the old behavior, or at least the
short logs that we want to trigger some backfill.
Fixes: https://tracker.ceph.com/issues/43810
Signed-off-by: Sage Weil <sage@redhat.com>
* refs/pull/32853/head:
qa/suites/fs/upgrade: finish at octopus
qa/suites/fs/upgrade: set min-compat-client to octopus
qa/suites/fs/upgrade: set pg_autoscale_mode=off after upgrade
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
* refs/pull/32750/head:
qa/suites/upgrade/*-x/stress-split: run latest python tests at end
qa/tasks/ceph_manager: kludge around /var/log/ceph permissions
mgr/pg_autoscaler: don't check anything until cluster is nautilus
qa/suites/upgrade: install python3-* as part of final upgrade step
qa/tasks/radosbench: only set object size if != block size
qa/tasks/ceph: simplify mon_health_to_clog suppression during restart
cls_hello: alias write_return_data -> writes_dont_return_data
ceph_test_cls_hello: only do returndata test on octopus+
qa: remove unnecessary package excludes in upgrades/nautilus*
qa: exclude cephadm from upgrade/nautilus-x
qa/suites/upgrade/mimic-x/parallel: fix msgr2 vs octopus ordering
qa/suites/upgrade/nautilus-x/stress-split: fix msgr2 vs octopus ordering
qa/suites/upgrade/mimic-x: fix msgr2 vs octopus ordering
qa/suites/upgrade/nautilus-x: end on octopus
qa/suites/upgrade/mimic-x: finish at octopus
qa/suites/upgrade/nautilus-x: disable TOO_FEW_PGS warning
qa/tasks/ceph: set mon_health_to_clog=false via mon config
qa/suites/upgrade/mimic-x: disable TOO_FEW_PGS warning
Reviewed-by: Kefu Chai <kchai@redhat.com>
* refs/pull/32788/head:
qa/tasks/mgr/dashboard: set pg_num to 32
mgr/pg_autoscaler: default to pg_num[_min] = 32
Reviewed-by: Sage Weil <sage@redhat.com>
The ceph.py task normally makes these permissive. But a package upgrade
can reset the permissions so that we can't read and write the temp
export files. (We put them in these dirs now because it's alreadly
mapped out of cephadm containers to the host.)
Signed-off-by: Sage Weil <sage@redhat.com>
These are new packages, so they won't install just by upgrading the old
packages, and they are needed for some of the tests.
Signed-off-by: Sage Weil <sage@redhat.com>
This is mostly pointless, except that the -O option for objects size
used to be -o for pre-octopus, so passing -O breaks the upgrade tests.
Fortunately, the upgrades use the defaults, so we can just skate by here.
Signed-off-by: Sage Weil <sage@redhat.com>
Building xfstests on CentOS 8 has build errors:
2020-01-23T21:15:14.939 INFO:tasks.workunit.client.0.smithi155.stderr:/home/ubuntu/cephtest/mnt.0/client.0/tmp/xfstests/src/fssum.c:616: undefined reference to `major'
2020-01-23T21:15:14.939 INFO:tasks.workunit.client.0.smithi155.stderr:/home/ubuntu/cephtest/mnt.0/client.0/tmp/xfstests/src/fssum.c:617: undefined reference to `minor'
2020-01-23T21:15:14.940 INFO:tasks.workunit.client.0.smithi155.stderr:collect2: error: ld returned 1 exit status
From: /ceph/teuthology-archive/pdonnell-2020-01-23_20:47:47-fs-wip-pdonnell-testing-20200122.210533-distro-basic-smithi/4697257/teuthology.log
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
These are needed for CentOS 8 / RHEL8.
Also organize the YAML to use the cleaner array syntax.
Fixes: https://tracker.ceph.com/issues/43459
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
* refs/pull/32787/head:
qa/tasks/cephadm: work around .git suffix on ceph_repo
qa/tasks/cephadm: learn to pull cephadm from github
Reviewed-by: Jan Fajerski <jfajerski@suse.com>
The dnsmasq package on centos 8.0 is broken, see
https://tracker.ceph.com/issues/43744
For now, run this test on ubuntu.
Signed-off-by: Sage Weil <sage@redhat.com>
* refs/pull/32726/head:
mon/ConfigMonitor: clean out bad config records
mon/ConfigMonitor: make 'config get' on fsid work
common/options: remove NO_MON_UPDATE from mon_dns_srv_name
mon/ConfigMonitor: do not 'config get' on NO_MON_UPDATE options
mon/ConfigMonitor: do not set NO_MON_UPDATE values
Reviewed-by: Neha Ojha <nojha@redhat.com>
Reviewed-by: Josh Durgin <jdurgin@redhat.com>
Using assert causes the tests to fail on teuthology due to this test being run
separately from the ceph cluster. Instead use stat for testing.
Signed-off-by: Varsha Rao <varao@redhat.com>
* refs/pull/31232/head:
test: test case for openfiletable MAX_ITEMS_PER_OBJ value verification
mds/OpenFileTable: match MAX_ITEMS_PER_OBJ to osd_deep_scrub_large_omap_object_key_threshold
Reviewed-by: Zheng Yan <zyan@redhat.com>
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
Rather than verify the latest OSDMap is the same as the rank's
osdmap_epoch_barrier, just use the rank's version. The OSDMap may change
out-of-band with the test startup and thus the epoch's would diverge.
The file system and rank is fresh for each test so there's no reason to
care if the MDS barrier is one epoch behind the latest.
Fixes: https://tracker.ceph.com/issues/43554
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
Make sure PGs peer (simply flushing state to mon isn't enough).
Fixes: https://tracker.ceph.com/issues/43721
Signed-off-by: Sage Weil <sage@redhat.com>
Instead of printing the (useless) traceback, just print a warning about
ignoring the failure. The traceback makes it harder to search for the
real problem in the teuthology log.
Fixes: https://tracker.ceph.com/issues/43718
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
* refs/pull/32524/head:
qa/xfstests_dev: change deps for xfstests-dev on ubuntu
qa/cephfs: change deps for xfstests-dev on centos8
vstart_runnner: add sh method to LocalRemote
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
These options are special. Since we can't get the values from the mon,
we also shouldn't store them there. Most of them are not options that
can be changed anyway.
Signed-off-by: Sage Weil <sage@redhat.com>
This was asserting that all PGs are active or peered, but that assertion
could fail if the concurrent workload created a new pool.
Switch to a loop that checks several times for the condition to be true.
Fixes: https://tracker.ceph.com/issues/43656
Signed-off-by: Sage Weil <sage@redhat.com>
The cores will make teuthology fail the job--and we don't want them for
this test, where we are deliberately causing crashes.
Fixes: https://tracker.ceph.com/issues/43653
Signed-off-by: Sage Weil <sage@redhat.com>
Rename python to python2 if Ubuntu distro release is 19 or later.
Fixes: https://tracker.ceph.com/issues/43522
Signed-off-by: Rishabh Dave <ridave@redhat.com>
* refs/pull/32713/head:
qa/tasks/cephadm: start watching initial daemons before bootstrap
qa/tasks/cephadm: create /etc/ceph if it doesn't exist
qa/tasks/cephadm: fix log whitelist when there is no whitelist
Reviewed-by: Josh Durgin <jdurgin@redhat.com>
Existing pools might have too many/few PGs and produce a warning that
prevents us from getting to HEALTH_OK.
Signed-off-by: Sage Weil <sage@redhat.com>
