Details
-
Bug
-
Resolution: Done-Errata
-
Blocker
-
None
-
RH-SSO-7.6.5
-
False
-
None
-
False
-
Description
Starting from RH SSO 7.6.5, all the Liveness and Readiness Probes timeouts if the CPU limits in OpenShift is any value below 3. When setting as 2, they might work for a while but after some time the pods are still restarted.
Before, customers were able to run RH SSO with 1 CPU or even less.
Please find below the Liveness Probes taking 55 seconds (way longer than the default 22 seconds timeout) to execute with 1 CPU limit:
sh-4.4$ time /probes/liveness_probe.sh
{
"outcome" => "success",
"result" => {
"name" => "<USERNAME>",
"attributes" => undefined
}
}
INFO Using the '<USERNAME>' username to authenticate the probe request against the JBoss DMR API.
{
"outcome" => "success",
"result" => {
"name" => "<USERNAME>",
"attributes" => undefined
}
}
INFO Using the '<USERNAME>' username to authenticate the probe request against the JBoss DMR API.
{
"outcome" => "success",
"result" => {
"name" => "<USERNAME>",
"attributes" => undefined
}
}
INFO Using the '<USERNAME>' username to authenticate the probe request against the JBoss DMR API.
{
"outcome" => "success",
"result" => {
"name" => "<USERNAME>",
"attributes" => undefined
}
}
INFO Using the '<USERNAME>' username to authenticate the probe request against the JBoss DMR API.
real 0m55.362s
user 0m24.459s
sys 0m2.363s
At least one customer reported they were able to run RH SSO until 7.6.4 with even 300ms of CPU limit before.
In addition to the data above, I will also attach the OpenShift even logs from my own local instance when also setting the CPU to 1.
Attachments
Issue Links
- duplicates
-
-
- Closed
-
- links to
-
RHBA-2023:120513
Red Hat Single Sign-On 7.6.5 for OpenShift image enhancement update
- mentioned on
