We are deploying containers manually. Mark them unmanaged so that we
do not fight against mgr/cephadm cleaning up orphan daemons.
Signed-off-by: Sage Weil <sage@redhat.com>
- For tests, use bleeding-edge octopus branch
- For production defaults, use ceph/ceph:v15.2 tag
- For bootstrap, grab cephadm script from latest octopus branch
Signed-off-by: Sage Weil <sage@redhat.com>
* refs/pull/33952/head:
qa/workunits/cephadm: --skip-mon-network when using 127.0.0.1
cephadm: add tests
qa/tasks/cephadm: pass -v to bootstrap
mgr/cephadm: only try to place mons on hosts matching public_network
mgr/cephadm: keep track of host networks, ips
cephadm: automatically infer mon public_network, if we can
cephadm: add list-networks command
Reviewed-by: Sebastian Wagner <swagner@suse.com>
This was present, but a no-op.
By default, install cephadm.
Optionally take a list of packages to install instead (e.g., ceph-common).
Signed-off-by: Sage Weil <sage@redhat.com>
This ticket seems to suggest that (1) the root cause is related to an
exec that is orphaned and screws up the container state (due to, e.g., ssh
dropping, or a timeout), (2) -f may be needed, sometimes, to recover, and
(3) newer versions fix it.
https://github.com/containers/libpod/issues/3226
Way back in 26f9fe54cb we found that using
-f the first time around was a Bad Idea, so we'd rather avoid this.
Instead, just avoid triggering the bug.
Signed-off-by: Sage Weil <sage@redhat.com>
docker returns '<no value>' if the label isn't present, in which case we
still need to run ceph -v.
Also, don't probe non-ceph (e.g., monitoring) containers.
Also, only probe each image id once.
Add a simple test.
Signed-off-by: Sage Weil <sage@redhat.com>
* refs/pull/33133/head:
qa/workunits/cephadm/test_cephadm.sh: make monitoring tests faster
qa/workunits/cephadm/test_cephadm: 2 OSDs is enough
cephadm: disable node-exporter cpu/memory limits for the time being
Reviewed-by: Michael Fritch <mfritch@suse.com>
2020-01-13T10:48:14.235 INFO:tasks.workunit.client.0.smithi177.stderr:+ vgchange -an test_cephadm
2020-01-13T10:48:14.239 INFO:tasks.workunit.client.0.smithi177.stderr: WARNING: Running as a non-root user. Functionality may be unavailable.
2020-01-13T10:48:14.239 INFO:tasks.workunit.client.0.smithi177.stderr: /run/lvm/lvmetad.socket: access failed: Permission denied
2020-01-13T10:48:14.240 INFO:tasks.workunit.client.0.smithi177.stderr: WARNING: Failed to connect to lvmetad. Falling back to device scanning.
2020-01-13T10:48:14.260 INFO:tasks.workunit.client.0.smithi177.stderr: /dev/mapper/control: open failed: Permission denied
2020-01-13T10:48:14.261 INFO:tasks.workunit.client.0.smithi177.stderr: Failure to communicate with kernel device-mapper driver.
2020-01-13T10:48:14.261 INFO:tasks.workunit.client.0.smithi177.stderr: Incompatible libdevmapper 1.02.145 (2017-11-03) and kernel driver (unknown version).
2020-01-13T10:48:14.301 INFO:tasks.workunit.client.0.smithi177.stderr: /run/lock/lvm/V_test_cephadm:aux: open failed: Permission denied
2020-01-13T10:48:14.301 INFO:tasks.workunit.client.0.smithi177.stderr: Can't get lock for test_cephadm
2020-01-13T10:48:14.301 INFO:tasks.workunit.client.0.smithi177.stderr: Cannot process volume group test_cephadm
Signed-off-by: Michael Fritch <mfritch@suse.com>
There's no need for mgr/cephadm to pass --mon-* arguments to
/usr/bin/cephadm just so it can add bits to the config file when the
config file is also being passed in from mgr/cephadm.
Signed-off-by: Sage Weil <sage@redhat.com>
Otherwise the test breaks if there are other cluster(s) on the host (like
there normally are in my dev environment).
Signed-off-by: Sage Weil <sage@redhat.com>
