This test requires that the Cray CLI is configured on nodes where the test is executed. See Cray command line interface.
This test can be run on any Kubernetes NCN (any master or worker NCN, but not the PIT node). When run on a Kubernetes master NCN, the TFTP file transfer subtest is omitted. However, that TFTP subtest is run on a worker NCN as part of the Goss NCN health checks.
(ncn-mw#
) The following command runs the entire SMS test suite (with the possible exception of the TFTP file
transfer subtest, as noted in the previous paragraph).
/usr/local/bin/cmsdev test -q all
cmsdev
tool logs to /opt/cray/tests/install/logs/cmsdev/cmsdev.log
/opt/cray/tests/cmsdev.log
cmsdev
resultsSUCCESS
.
SUCCESS: All 7 service tests passed: bos, cfs, conman, crus, ims, tftp, vcs
FAILURE
and lists the failed checks.
FAILURE: 2 service tests FAILED (conman, ims), 5 passed (bos, cfs, crus, tftp, vcs)
all
in the cmsdev
command line with the name of the service. For example: /usr/local/bin/cmsdev test -q cfs
Additional test execution details can be found in /opt/cray/tests/install/logs/cmsdev/cmsdev.log
.
Some of the subtests may fail if the Cray CLI is not configured on the management NCN where cmsdev
is executed.
See the following for more information:
In CSM 1.4.0, the CRUS subtest of cmsdev
may fail with an error resembling the following:
ERROR (run tag KPEqc-crus): CLI command failed (and does not look like a CLI config issue) (crus session list --format json)
This is because of a known issue in CSM 1.4.0 that is fixed in CSM 1.4.1. For more information, see CRUS Subcommands Missing From Cray CLI.
If an Etcd restore has been performed on one of the SMS services (such as BOS or CRUS), then the first Etcd pod that comes up after the restore will not have a PVC (Persistent Volume Claim) attached to it (until the pod is restarted). The Etcd cluster is in a healthy state at this point, but the SMS health checks will detect the above condition and may report test failures similar to the following:
ERROR (run tag 1khv7-bos): persistentvolumeclaims "cray-bos-etcd-ncchqgnczg" not found
ERROR (run tag 1khv7-crus): persistentvolumeclaims "cray-crus-etcd-ffmszl7bvh" not found
In this case, these errors can be ignored, or the pod with the same name as the PVC mentioned in the output can be restarted (as long as the other two Etcd pods are healthy).
On systems where too many BOS v1 sessions exist, the cmsdev
test will hang when trying to
list BOS v1 sessions. See Hang Listing BOS V1 Sessions for more
information.