Launchpad itself

supermirror latency monitoring

Bug #44205 reported by David Allouche
14
Affects Status Importance Assigned to Milestone
Launchpad itself
Fix Released
Medium
Unassigned

Bug Description

I updated the bzrk mainline branch on monday, but it took more than 12 hours before being published (that's an unacceptable latency), and the supermirror-pull script did not send any message to the launchpad-errors@ mailing list.

There is not enough information to diagnose latency issues with the branch puller script. It should log to a file (with timestamps) when it starts, which branches it processes (by filesystem id and unique_name) the errors (even non-fatal) it encounters, and when it finishes.

The log files should probably be rotated and deleted after some time to prevent the log data to grow without bound.

Tags: lp-code
Revision history for this message
Tim Penhey (thumper) wrote :

David, is this still valid?

Changed in launchpad-bazaar:
assignee: nobody → ddaa
Revision history for this message
David Allouche (ddaa) wrote :

Yes, we do to have any automatic way to measure the performance of the supermirro at the moment. The current latency measurement is based on errormail caused by script lock contention.

Maybe this should be implemented by recording times into the database instead of lockfile, but this work is still to be done.

Revision history for this message
David Allouche (ddaa) wrote :

> Maybe this should be implemented by recording times into the database instead of lockfile, but this work is still to be done.

I meant "instead of LOG FILE" not lockfile.

Revision history for this message
David Allouche (ddaa) wrote :

Unassigning bug, because I do not see myself working on it in the short term.

Changed in launchpad-bazaar:
assignee: ddaa → nobody
status: Unconfirmed → Confirmed
Revision history for this message
Michael Hudson-Doyle (mwhudson) wrote :

So I'm going to work on reporting supermirror script runtimes in the db, which isn't quite what this bug is after, but it's as close as we're like to get in the near term...

Changed in launchpad-bazaar:
assignee: nobody → mwhudson
Revision history for this message
Jonathan Lange (jml) wrote :

To be honest, I'm tempted to mark this as done. I don't see how further work would give us any benefit.

It *might* be nice to have a cricket graph of mirroring runtimes, I guess, but that would be an RT ticket.

Revision history for this message
Michael Hudson-Doyle (mwhudson) wrote :

Unassigning, and marking done: I can't see any more work that remains to be done in this area.

Changed in launchpad-bazaar:
assignee: mwhudson → nobody
status: Confirmed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Duplicates of this bug

Subscribing...

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.