Created attachment 1399435 [details] list of updates applied Description of problem: The .all alias exists but after a certain date, it is longer updated - even though other indices _are_ created and are searchable after this date. This date coincides with platform updates the customer applied (see attached for a list) and then a reboot. OCP was a fresh install and not an upgrade. Version-Release number of selected component (if applicable): OCP 3.6.173.0.5 Dockerfile info: /root/buildinfo/Dockerfile-openshift3-logging-curator-v3.6.173.0.96-2 "build-date"="2018-01-10T22:14:06.235392" Dockerfile info: /root/buildinfo/Dockerfile-openshift3-logging-elasticsearch-v3.6.173.0.96-3 "build-date"="2018-01-18T13:02:43.616483" Dockerfile info: /root/buildinfo/Dockerfile-openshift3-logging-auth-proxy-v3.6.173.0.96-2 "build-date"="2018-01-10T22:14:09.238369" Dockerfile info: /root/buildinfo/Dockerfile-openshift3-logging-fluentd-v3.6.173.0.96-2 "build-date"="2018-01-10T22:14:59.119836" How reproducible: Not certain. The .all alias stopped being updated after the customer applied platform updates and rebooted (see attached for list of updates applied). Steps to Reproduce: 1.install OCP 3.6.173.0.5 2.apply updates listed in attached 3.reboot platform Actual results: .all alias stops updating Expected results: .all alias doesn't stop updating and continues to be searchable Additional info: Customer also has a 3.7 cluster that may be exhibiting the same problems. logging-dump.sh script attached.
https://github.com/openshift/origin-aggregated-logging/pull/973
No. This will alias the indices and make no modification to the data. The customer will need to manually alias old indices
Can you advise if these indices were created before or after the deployment of the new images? The change exists [1] in the referenced version but will only apply to new indices http://pkgs.devel.redhat.com/cgit/rpms/logging-elasticsearch-docker/tree/index_templates/com.redhat.viaq-openshift-operations.template.json?h=rhaos-3.7-rhel-7
These indices were created after the deployment of the new images. The test env was a new env and I didn't use dynamic pv.
Suresh, we will not be backporting to 3.5. Consider upgrading to a newer release. Alternatively you could modify the index-templates to add the alias upon each index creation https://github.com/openshift/origin-aggregated-logging/pull/1230
The .all alias was added when index is created in v3.7.57-4.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2018:2234