Skip to content
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: degrade a log message to level 2 #6929

Merged
merged 1 commit into from Dec 17, 2015
Merged

Conversation

clever215
Copy link
Contributor

FIX #14083

When leader election was done successfully, it will ask peon to collect info. When failed, it will cause monitor to do reelection, which makes relevant log message so important.
So It would be better if reducing the log level to a default value.

Signed-off-by: Kongming Wu wu.kongming@h3c.com

FIX ceph#14083

When leader election was done successfully, it will ask peon to collect info. When failed, it will cause monitor to do reelection, which makes relevant log message so important.
So It would be better if reducing the log level from 5 to 2.

Signed-off-by: Kongming Wu <wu.kongming@h3c.com>
@ghost ghost added cleanup core labels Dec 15, 2015
@jecluis
Copy link
Member

jecluis commented Dec 15, 2015

I think logging this at level 0 may cause unnecessary panic, especially given this is normal behavior.

Logging it at level 1 would be better than logging at level 0. That way you can simply set 'debug paxos = 1' to catch it, without having the noise from other messages from higher levels (such as 5). Other users will likely not care about this as long as the monitors join the quorum at some point; if they do, they too can set 'debug paxos = 1' to catch it.

@clever215
Copy link
Contributor Author

@jecluis OK! That sounds sensible. I've done updating.

@clever215 clever215 changed the title mon: degrade a log message to a default level mon: degrade a log message to level 1 Dec 16, 2015
@jecluis
Copy link
Member

jecluis commented Dec 16, 2015

@clever215 I know this is annoying, but I just realized the default log level for paxos is '1'. And that's likely why this is set to '5' to begin with. I'd ask you to bear with me and set '2' instead of '1'. The reasoning is the same I gave before. Sorry about that!

@clever215
Copy link
Contributor Author

@jecluis I just discussed with my colleague, and she said it will not product this log message so many times if monitor election is running without error. Only DB broken that she happened, and She found this log message is convenient to locate issue. But what you said is also reasonable.

@jecluis
Copy link
Member

jecluis commented Dec 16, 2015

What your colleague is seeing is not because the store is broken. Most likely, the store being "broken" is enhancing the reasons why this happens (slow monitor, falling behind etc).

If you want to easily reproduce this state, you just have to keep a monitor out of quorum for long enough that, by the time it tries to join, the state has advanced enough that you do not share any paxos versions.

IIRC, this will happen whenever a new monitor is joining the quorum, for instance.

@clever215 clever215 changed the title mon: degrade a log message to level 1 mon: degrade a log message to level 2 Dec 16, 2015
@clever215
Copy link
Contributor Author

@jecluis Yes, you are right!

jecluis added a commit that referenced this pull request Dec 17, 2015
mon: degrade a log message to level 2

Reviewed-by: Joao Eduardo Luis <joao@suse.de>
@jecluis jecluis merged commit cf5da01 into ceph:master Dec 17, 2015
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
3 participants