Fixed
Pinned fields
Click on the next to a field label to start pinning.
Details
Assignee
Apaar GuptaApaar GuptaReporter
Apaar GuptaApaar GuptaIs this a Regression?
UnknownTriage
UntriagedDue date
Jan 22, 2024Story Points
0Priority
MajorInstabug
Open Instabug
Details
Details
Assignee
Apaar Gupta
Apaar GuptaReporter
Apaar Gupta
Apaar GuptaIs this a Regression?
Unknown
Triage
Untriaged
Due date
Jan 22, 2024
Story Points
0
Priority
Instabug
Open Instabug
PagerDuty
PagerDuty
PagerDuty
Sentry
Sentry
Sentry
Zendesk Support
Zendesk Support
Zendesk Support
Created January 18, 2024 at 4:44 PM
Updated September 12, 2024 at 6:12 PM
Resolved January 22, 2024 at 4:48 PM
Supportal faced downtime caused by their three-node 7.1.2 couchbase cluster. The Magma bucket was flooding the logs and constantly retrying a file delete on one of their kv nodes as reported possibly causing the slowness and preventing debugging of the issue.
2024-01-18T10:59:30.802799+00:00 WARNING [(default) magma_5/kvstore-773/rev-000000002/localIndex]Failed to delete state file 5972 reason:IOError: /mnt/couchbase/data/default/magma.5/kvstore-773/rev-000000002/localIndex/state.5972: Unable to delete : File not found err=No such file or directory. Closest non-empty parent directory: /mnt/couchbase/data/default/magma.5/kvstore-773/rev-000000002/localIndex offset:0 len:0
This issue was probably caused by an intermittent filesystem issue resulting in a file remove/unlink returning a bad status despite the file being deleted causing Magma to get stuck in a retry loop trying to delete the file. Rebooting the node fixed the issue. A defensive fix is required to prevent Magma from flooding the logs and potentially pinning a CPU core causing slowness.