DELETE Cluster Ops Team
[DCOS_OSS-1486] Metrics agent crashes when the mesos containers endpoint is missing fields Created: 31/Jul/17 Updated: 09/Nov/18 Resolved: 02/Aug/17 |
|
| Status: | Resolved |
| Project: | DC/OS |
| Component/s: | dcos-metrics |
| Affects Version/s: | DC/OS 1.9.1, DC/OS 1.9.2 |
| Fix Version/s: | DC/OS 1.9.3, DC/OS 1.10.0 |
| Type: | Bug | Priority: | Blocker |
| Reporter: | Philip Norman (Inactive) | Assignee: | Philip Norman (Inactive) |
| Resolution: | Done | ||
| Labels: | beta2, clusterops, clusterops-team | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Team: | DELETE Cluster Ops Team |
| Sprint: | ClusterOps Sprint 1.11-3 |
| Story Points: | 5 |
| Description |
|
dcos-metrics polls the Mesos agent /containers endpoint and stores the contents of the statistics object in each item. It appears that under some circumstances, the /containers items don't have a statistics object; the dcos-metrics code expects it always to be present and this is causing a segfault which crashes the whole process. |
| Comments |
| Comment by Philip Norman (Inactive) [ 31/Jul/17 ] |
| Comment by Philip Norman (Inactive) [ 07/Aug/17 ] |
| Comment by Jan-Philip Gehrcke (Inactive) [ 16/Aug/17 ] |
|
This was linked from https://github.com/dcos/dcos/pull/1810 which I just merged via train 203 (https://github.com/dcos/dcos/pull/1831). I expected to be able to close a corresponding JIRA ticket with fixVersion 1.10. This here has fixVersion 1.10, but it has already been closed before :/. |