Update logging workload to OLO 5.6 #549
Conversation
| Storage class to use for the persistent storage. The faster the storage, better the Elasticsearch performance. | ||
|
|
||
| ### ES_STORAGE_SIZE | ||
| Default: `100G` | ||
| Each data node in the cluster is bound to a Persistent Volume Claim that requests the size specified using this variable from the cloud storage. | ||
|
|
||
| ### ES_MEMORY_LIMITS | ||
| Default: `16Gi` |
There was a problem hiding this comment.
It looks like the default request and limit are the same. Do we really even need to set a limit here?
There was a problem hiding this comment.
File comes from the docs (https://docs.openshift.com/container-platform/4.12/logging/cluster-logging-deploying.html#cluster-logging-deploy-cli_cluster-logging-deploying), but yes we can remove it if we want
There was a problem hiding this comment.
We can keep this variable as long as defaults make sense, which is the case :)
There was a problem hiding this comment.
Agree, request=limit makes this a guaranteed pod which has implications in the context of preemption - guaranteed pods are preempted last, so there's probably several reasons why we have requests=limit for ES.
There was a problem hiding this comment.
I don't disagree that the values may be reasonable, etc. My thinking here was if we are testing the logging stack we shouldn't limit the backend data store as it is just something that we may hit (OOM of ES) which would negatively impact our tests but not be prudent to the results. I.e. if we cause ES to OOM it does NOT indicate a failed test. It just means we need to change those values and re-run. IMHO removing the limit just potentially saves us steps/debugging down the line.
morenod
force-pushed
the
logging_hypershift
branch
from
f493509 to
7603128
Compare
| @@ -44,7 +44,7 @@ function cleanup() { | |||
| while [[ $( oc get projects | grep -w "openshift-logging\|openshift-operators-redhat") ]]; do | |||
| sleep 5 | |||
| wait_time=$((wait_time+5)) | |||
| if [[ $wait_time -ge $TIMEOUT ]]; then | |||
| if [[ "${wait_time}" -ge "${TIMEOUT}" ]]; then | |||
There was a problem hiding this comment.
Can we can reduce this whole function into:
cleanup() {
oc delete --wait=true project openshift-logging openshift-operators-redhat --ignore-not-found --timeout=${TIMEOUT}
}
There was a problem hiding this comment.
rewritted cleanup function to use the --wait
morenod
force-pushed
the
logging_hypershift
branch
3 times, most recently
from
20178f5 to
3400d08
Compare
morenod
force-pushed
the
logging_hypershift
branch
from
3400d08 to
6f9b6aa
Compare
morenod
force-pushed
the
logging_hypershift
branch
from
6f9b6aa to
6410a94
Compare
morenod
force-pushed
the
logging_hypershift
branch
from
6410a94 to
c5443bd
Compare
Update code of logging workload to the docs published for 4.12 and OLO 5.6