Bug 1430484

Summary: Upgrade from 3.3 to 3.4, Insufficient Pods
Product: OpenShift Container Platform Reporter: Steven Walter <stwalter>
Component: NodeAssignee: Seth Jennings <sjenning>
Status: CLOSED ERRATA QA Contact: Anping Li <anli>
Severity: high Docs Contact:
Priority: unspecified    
Version: 3.4.1CC: aos-bugs, decarr, erich, jokerman, mmccomas, wmeng
Target Milestone: ---   
Target Release: 3.7.0   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: No Doc Update
Doc Text:
undefined
Story Points: ---
Clone Of: Environment:
Last Closed: 2017-11-28 21:53:01 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:

Description Steven Walter 2017-03-08 18:01:05 UTC
Description of problem:
Getting "Insufficient Pods" after upgrading to 3.4, causing a large number of pods to suddenly stop working. This is because of the change from a flat max-pods, to whichever is smaller of {250,10pods/core}. However this is not yet documented (https://bugzilla.redhat.com/show_bug.cgi?id=1376585) and post-upgrade, it is unknown how to resolve this. In existing clusters that have many (small) pods, after upgrade suddenly the number of pods is higher than the pods-per-core mandate, causing dozens of pods to be unschedulable. It appears from looking around that the pods-per-core number can be configured in ansible hosts, but there is no documentation so it is hard to verify this, or what steps to take to increase this number.

Version-Release number of selected component (if applicable):
3.4.1


Marking as high severity given this can cause an outage to applications and we do not warn of this.

Additional info:

https://github.com/openshift/openshift-docs/issues/2455
https://bugzilla.redhat.com/show_bug.cgi?id=1371309

Comment 5 Seth Jennings 2017-03-10 17:14:00 UTC
Origin Docs PR:
https://github.com/openshift/openshift-docs/pull/3916

As far as resolving the situation post-upgrade, the only thing to do is to set the pods-per-core >= max-pods so that max-pods becomes the limiting factor.

This can be done in the installer inventory file:
openshift_node_kubelet_args={'pods-per-core': ['200'], 'max-pods': ['200']}

Comment 6 Anping Li 2017-03-22 10:06:09 UTC
Specified openshift_node_kubelet_args={'pods-per-core': ['200'], 'max-pods': ['200']} can limit the pod number, the pod can be limited by the configuration.  so move bug to verified.

Comment 11 errata-xmlrpc 2017-11-28 21:53:01 UTC
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/RHSA-2017:3188