process instalbility observed when running inside a container laucnhed with daemonset

[sbezverk]

I noticed some issue when launching container using the daemonset. The same container launched with ReplicationController works 100% stable. The issue are mostly related to a process restart, some errors and inability to create a named socket. It is not a config issue as, as soon as I add strace in the command line of the process, it starts working fine. It seems some sort of a racing condition with daemoset. The issue is 100% reproducible, I am ready to offer access to the impacted test bed for further troubleshooting. Here is the error which is generated when container starts with DaemonSet:
[19202.800318] traps: ovsdb-server[24147] general protection ip:7fb18600be37 sp:7ffef3da0d00 error:0 in libc-2.17.so[7fb185fd5000+1b7000]
I am on Cenots 7.2, kubernetes 1.3.4-dirty (I needed this version for another issue which is fixed in this release). Please let me know if somebody would be interested to check it out.

[dchen1107]

@sbezverk If possible, could you please provide the following information:

1. Run your container through rc, then run command kubectl get -o yaml pods
2. Run your container as daemonset, then run command kubectl get -o yaml pods
3. Run your container as daemonset, run kubectl logs --previous pod-id container to retrieve previous terminated container's log

[sbezverk]

@dchen1107 It turned out that replication controller also shows the similar issue :(
Here is the link to requested information.
Collected logs
Let me know if additional info is required.

[sbezverk]

@dchen1107 Interesting observation, if I start the process inside of a container with this script,
#!/bin/bash

sleep 1
/usr/sbin/ovsdb-server /etc/openvswitch/conf.db -vconsole:emer -vsyslog:err --remote=punix:/run/openvswitch/db.sock -vfile:info --log-file=/var/log/kolla/openvswitch/ovsdb-server.log

it works perfectly on both compute nodes.

[fejta-bot]

Issues go stale after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

Prevent issues from auto-closing with an /lifecycle frozen comment.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or @fejta.
/lifecycle stale

[fejta-bot]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or @fejta.
/lifecycle rotten
/remove-lifecycle stale

[fejta-bot]

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close