https://bugzilla.wikimedia.org/show_bug.cgi?id=70752

--- Comment #2 from Bryan Davis <[email protected]> ---
Was this the host where Roan killed the ssh session from tin or another host?
In general this looks like rsync experiencing some sort of major on the
receiving client side. With the horrible sync times and extreme load that the
mw1010 server was under during this scap I'm actually not surprised to see a
few errors like this occur.

I'm not sure there is much that could be done to adjust for this in the scap
program itself. We could dial back the number of parallel processes we run or
try to add some sort of load check in the rsync server selection process. The
settings we are running now however have remained the same for many months and
we have only seen these really bad sync times and high load on the rsync
servers in the last week or two. Many (all?) of the rsync servers are job
runner nodes. I'm sort of wondering if something has changed in the behavior or
job load that is clashing with scap.

I'd really like to move from rsync to a git fetch or other method of
synchronization that uses precomputed diffs and lessens the computational cost
on both the client and server.

-- 
You are receiving this mail because:
You are the assignee for the bug.
You are on the CC list for the bug.
_______________________________________________
Wikibugs-l mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/wikibugs-l

Reply via email to