Actions
Bug #53749
openceph device scrape-health-metrics truncates smartctl/nvme output to 100 KiB
Status:
New
Priority:
Normal
Assignee:
-
Category:
-
Target version:
-
% Done:
0%
Source:
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Component(RADOS):
Pull request ID:
Crash signature (v1):
Crash signature (v2):
Description
- https://github.com/ceph/ceph/blob/ae17c0a0c319c42d822e4618fd0d1c52c9b07ed1/src/common/blkdev.cc#L729
- https://github.com/ceph/ceph/blob/ae17c0a0c319c42d822e4618fd0d1c52c9b07ed1/src/common/blkdev.cc#L659
If I'm not mistaken, smartctl output can easily be longer than 100 KiB, especially when there are errors in the log (and that's when it matters).
I understand the motivation to provision against the output being arbitrarily large, but I think it would be wise to detect that situation (noticing that we're not at EOF of `get_stdout()`), and record that somehow, so that the user can notice that they are looking at truncated output, instead of the output just stopping at some point, in the worst case without the user noticing.
No data to display
Actions