| Summary: | flannel system container - service fails to start after reboot on rhel atomic host | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 7 | Reporter: | Michael Nguyen <mnguyen> |
| Component: | flannel-container | Assignee: | Avesh Agarwal <avagarwa> |
| Status: | CLOSED ERRATA | QA Contact: | Martin Jenner <mjenner> |
| Severity: | high | Docs Contact: | |
| Priority: | unspecified | ||
| Version: | 7.3 | CC: | avagarwa, mnguyen, ypu |
| Target Milestone: | rc | Keywords: | Extras |
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | flannel-docker-0.5.5-5 | Doc Type: | If docs needed, set a value |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2016-11-04 09:11:37 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: | |
Let me know more details here: 1. why are you using this image gscrivano/etcd? why not official rhel/atomic's etcd image? 2. how did you install flannel? is it rpm or container? 1. I just tried with the official rhel/atomic etcd image and maybe hitting this issue: https://bugzilla.redhat.com/show_bug.cgi?id=1376123 The service won't start after using atomic install --system to install the container. 2. I am installing flannel as a container. 1. Even with the issue, etcd container service will start without any issue except that systemctl will time out. For a workaround, to avoid that after installation, just remove Type=notify manually from the service unit file, and then start the service. 2. Could you please tell the exact command you used for flannel installation. 3. Do you have flannel or etcd rpms installed on the machine? To find out if they are interfering with the container installation? 1. I removed Type=notify and the containers start without the timeout now but am still seeing the restart issue after reboot. 2. atomic install --system --name=flannel brew-pulp-docker01.web.prod.ext.phx2.redhat.com:8888/rhel7/flannel 3. I just tried removing flannel and etcd rpms then installing the containers. I am seeing the same issue with flannel starting after reboot. The issue is reproducible and the cause is that after reboot /etc/systemd/system/docker.service.d/flannel.conf and /run/flannel are deleted and not recreated during next restart of flannel. I have noticed that at times, even during the initial start, these files are not created and flannel fails. I think I have found the root cause and the problem is that bind mount of /etc/systemd/system/docker.service.d/ is incorrect and point to a file flannel.conf. I have pushed an update to its upstream repo here: https://github.com/aveshagarwal/flannel-container/commit/1a177a55496af0fba4b0ac6d7d3e80a1bbd15133 I will push this change to rhel/atomic in the next rebuild for rc candidate. I've tested flannel-docker-0.5.5-5 and the service starts after reboot now. Test with the newest version of flannel-docker-0.5.5-6 with the steps in comment #1. And seems after the system rebooted. Flannel and docker all works fine. So set this to verified. 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-2016:2650 |
Description of problem: flannel container fails to start after reboot on rhel atomic host Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. atomic install --system --name=etcd gscrivano/etcd 2. systemctl start etcd 3. runc exec etcd etcdctl set /atomic.io/network/config '{"Network":"172.17.0.0/16"}' 4. systemctl start flannel 5. systemctl daemon-reload 6. systemctl reboot 7. Check status of flannel and docker Actual results: Flannel is not started and prevents docker from restarting Expected results: Flannel and docker services are up and running Additional info: