persistent failure of atomic kv operations until docker is restarted

[alfredr]

Frequently, I get errors of the following form on container start which persist until docker is restarted.

docker start e834638cd1db -- Error response from daemon: Cannot start container e834638cd1dbbc60209a772c385d5f15d8a86b8d12a96e04d077e52869555b0c: Unable to complete atomic operation, key modified

Other docker hosts which use the same consul instance as a kv store for libnetwork do not have this issue and consul continues to operate normally.

[GordonTheTurtle]

@alfredr It has been detected that this issue has not received any activity in over 6 months. Can you please let us know if it is still relevant:

• For a bug: do you still experience the issue with the latest version?
• For a feature request: was your request appropriately answered in a later version?

Thank you!
This issue will be automatically closed in 1 week unless it is commented on.
For more information please refer to #1926

[fentas]

had the same issue right now with Docker version 17.10.0-ce, build f4ffd25.
service docker restart helped. 👍

But I guess unresolved issue.

[abhi]

@fcrisciani they are not related. Lets track it with moby issues.

[fcrisciani]

@abhi the description is actually fitting perfectly the same behavior that I saw, but it's ok to track it on moby, maybe you can post here the reference issue to have a link.

[mavenugo]

@fcrisciani i opened moby/moby#35310 to track this. This is unrelated to the IPAM changes. But the endpoint_cnt states are messed up due to the lb-endpoint management.

[fcrisciani]

@mavenugo sure no prob, agree to keep only one place to track, I previously reopened this one because was generic enough to track the error message experienced.