Bug 1468845
Summary: | logging-fluentd pod segfaulting and restarting in some nodes - configured for logging-mux secure forward | ||
---|---|---|---|
Product: | OpenShift Container Platform | Reporter: | Mike Fiedler <mifiedle> |
Component: | Logging | Assignee: | Rich Megginson <rmeggins> |
Status: | CLOSED WORKSFORME | QA Contact: | Mike Fiedler <mifiedle> |
Severity: | medium | Docs Contact: | |
Priority: | unspecified | ||
Version: | 3.6.0 | CC: | aos-bugs, jcantril, mifiedle, nhosoi, pportant, rmeggins, xtian |
Target Milestone: | --- | ||
Target Release: | --- | ||
Hardware: | x86_64 | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | If docs needed, set a value | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2017-07-18 16:23:35 UTC | Type: | Bug |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: |
Description
Mike Fiedler
2017-07-08 20:50:55 UTC
Is this a test blocker? If so, please raise the priority and/or severity. I've recently updated the fluentd image to have a lot of new packages, among them new versions of fluentd itself. openshift3/logging-fluentd:v3.6.140-2 - please try the new image Also, do you see the core dumps correlated with either fluentd running OOM, or unable to connect to mux? I've moved up to 3.6.140-1 - does it have the new fluentd? 140-2 is not yet available on the mirror. (In reply to Mike Fiedler from comment #4) > I've moved up to 3.6.140-1 - does it have the new fluentd? 140-2 is not > yet available on the mirror. I built v3.6.140-2 this morning - not sure how long it takes to show up in the mirrors. You definitely need v3.6.140-2 - not sure if this image is also tagged v3.6.140 - I suppose you could docker pull both and see if they are the same. I think you need to rebuild the logging-mux pod with the new fluentd too. (In reply to Mike Fiedler from comment #6) > I think you need to rebuild the logging-mux pod with the new fluentd too. How? The mux pod and the fluentd daemonset use the same fluentd image. Disregard comment 6 - it uses logging-fluentd in the dc When I start logging-mux with the new logging-fluentd container image I get: oc logs logging-mux-11-kktns │Status: Downloaded newer image for brew-pulp-docker01.web.prod.ext.phx2.redhat.com:8888/openshift3/logging run.sh: line 108: bc: command not found The logging-mux pod goes into CrashLoopBackoff (In reply to Mike Fiedler from comment #9) > The logging-mux pod goes into CrashLoopBackoff The origin Dockefile has this addition to "yum install" list. I'd think it's needed for the 3.6 build... --- a/fluentd/Dockerfile +++ b/fluentd/Dockerfile @@ -23,6 +23,7 @@ RUN rpmkeys --import file:///etc/pki/rpm-gpg/RPM-GPG-KEY-CentOS-7 && \ ruby-devel \ libcurl-devel \ make \ + bc \ yep - rebuilding done - try v3.6.140-3 Mike, is this still happening with the latest image? Can we close this? Did not see this in later runs. Changing status to CLOSED->WORKSFORME since no fix delivered. |