Bug 1368641
Summary: | Memory leak in swift used as backend to glance | |||
---|---|---|---|---|
Product: | Red Hat OpenStack | Reporter: | Robin Cernin <rcernin> | |
Component: | openstack-tripleo-heat-templates | Assignee: | Christian Schwede (cschwede) <cschwede> | |
Status: | CLOSED CURRENTRELEASE | QA Contact: | Arik Chernetsky <achernet> | |
Severity: | high | Docs Contact: | ||
Priority: | high | |||
Version: | 8.0 (Liberty) | CC: | aschultz, cschwede, derekh, egafford, ekuvaja, jmelvin, ltoscano, mandreou, markmc, mburns, pgrist, rhel-osp-director-maint, smykhail, srevivo, zaitcev | |
Target Milestone: | async | Keywords: | Triaged | |
Target Release: | 8.0 (Liberty) | |||
Hardware: | Unspecified | |||
OS: | Unspecified | |||
Whiteboard: | ||||
Fixed In Version: | openstack-tripleo-heat-templates-0.8.14-34.el7ost | Doc Type: | If docs needed, set a value | |
Doc Text: | Story Points: | --- | ||
Clone Of: | ||||
: | 1533948 (view as bug list) | Environment: | ||
Last Closed: | 2018-09-25 08:57:40 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: | ||||
Bug Depends On: | ||||
Bug Blocks: | 1533948 |
Description
Robin Cernin
2016-08-20 09:44:28 UTC
I'm pretty sure this is related to a missing proxy-logging middleware in the beginning of the pipeline in /etc/swift/proxy-server.conf Could you share that file here? This issue has been reported earlier on in the CI: https://bugs.launchpad.net/tripleo/+bug/1595916 And a fix was proposed too: https://review.openstack.org/#/c/340389/ I was not able to reproduce this with a Swift-only installation. Thanks for sharing! You need the proxy-logging middleware twice, once at the beginning and once at the end. The correct one would look like this: pipeline = catch_errors healthcheck proxy-logging cache ratelimit tempurl formpost authtoken keystone staticweb proxy-logging proxy-server Some background on this: if an error happens before the end of the pipeline, it will not be properly logged if there is no proxy-logging at the beginning of the pipeline. The fix for director is merged upstream: https://review.openstack.org/#/c/340389/1/puppet/hieradata/controller.yaml I think we need a backport for this. Moving to openstack-tripleo-heat-templates Back to Christian backport the change for OSP8. >>>>>> Christian's reply: Please don't remove the proxy-logging middleware, but add it a second time to the beginning of the pipeline after health_check - removing will most likely break logging for the customer (pls see also my comment https://bugzilla.redhat.com/show_bug.cgi?id=1368641#c8). FTR - I already had a quick look into this, and the issue goes like this: some non-critical warning is logged in some of the middlewares before the end of the pipeline (where the proxy-logging is added in OSP7/8), and that breaks because logging is not yet defined. Therefore you need a second proxy-logging middleware early on in the pipeline. I will have a closer look next week to the issue itself; anyways the correct way is to always define the proxy-logging twice in the pipeline - at least that's how Swift has to be configured for a very long time. <<<<< |