New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
mon/PGMap: remove or narrow columns 'pg ls' output #20945
Conversation
The timestamps are more informative. Signed-off-by: Sage Weil <sage@redhat.com>
Degraded and misplaces are the more useful values (and the heading was too wide!). Signed-off-by: Sage Weil <sage@redhat.com>
Having two "PRI" columns with the same name is a bit weird, I wonder if we could remove the need for those columns at all with a convention that the OSDs in the UP and ACTING columns always have the primary first, if they don't already? |
It's normally true for replicated PGs, but not for EC ones. The operator<< in the OSD debug code print something like "[1,2,3]p2" (for ec pool soly)... we could use the same syntax here and stick it in the up/acting columns? It's a bit less clean but more compact. |
Since this output is already pretty low level stuff, that compact form should be a good fit. If we have two columns with the same name there is a risk it might bug me forever :-D |
Signed-off-by: Sage Weil <sage@redhat.com>
e.g., "[1,2,3]p1". This matches what PG's operator<< does in the OSD debug log, with the exception that the OSD leaves off the 'p123' bit on replicated pools (which currently always make the first acting osd the primary). Leaving it on here to be explicit and a bit future-proof. Signed-off-by: Sage Weil <sage@redhat.com>
retest this please |
@jcsp ping |
I think the removals are uncontroversial.
I'm not sure about the duplicate PRI columns, though. Perhaps we don't need up_primary at all?