[SERVER-32892] Unexpected Shutdown Version 3.6 Created: 25/Jan/18 Updated: 24/Feb/19 Resolved: 30/Jan/18 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | None |
| Affects Version/s: | 3.6.2 |
| Fix Version/s: | None |
| Type: | Question | Priority: | Major - P3 |
| Reporter: | Juan Soto (Inactive) | Assignee: | Mark Agarunov |
| Resolution: | Done | Votes: | 0 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Participants: | |
| Case: | (copied to CRM) |
| Description |
|
Hello, When I startup my mongod used as oplog store (Ops Manager), it dies.
Thanks |
| Comments |
| Comment by Mark Agarunov [ 30/Jan/18 ] | ||
|
I'm glad to hear it worked! Note that you can also set these values more permanently by copying the parameters returned by the first systemctl command into the [service] section of the mongod.service file with modified values. | ||
| Comment by Juan Soto (Inactive) [ 30/Jan/18 ] | ||
|
You are right. It was fixed!! | ||
| Comment by Mark Agarunov [ 25/Jan/18 ] | ||
|
juan.soto, to verify Bruce's timeout theory, and assuming this is using systemd as it is rhel7, could you run
to see if the values match the 90 seconds? If the values do match, you can confirm this is systemd killing the process by adding --no-block to the service startup:
This should prevent systemd from watching/waiting for the process. | ||
| Comment by Bruce Lucas (Inactive) [ 25/Jan/18 ] | ||
|
The SIGTERM was received almost exactly 90 seconds after the first log entry, while mongod was still starting up. Is it possible that an external monitoring process timed out and attempted to shut mongod down because it hadn't completed startup within 90 seconds? The startup was running in the oplog scanning phase when this occurred. We have seen this phase be particularly slow when mongod is started on an unwarmed volume restore from an AWS snapshot, because the initial data access in this case can be very slow. | ||
| Comment by Juan Soto (Inactive) [ 25/Jan/18 ] | ||
|
Thanks Dan. But the issue is that if we delete all the files of the database, the mongod is running ok. | ||
| Comment by Daniel Pasette (Inactive) [ 25/Jan/18 ] | ||
|
Thanks for the report Juan. It looks like the process got a shutdown signal (SIGTERM):
This is external to the mongod process. |