https://bugzilla.wikimedia.org/show_bug.cgi?id=61882

--- Comment #2 from Gabriel Wicke <gwi...@wikimedia.org> ---
My reply:

Without batching this is taking down the Parsoid cluster for at least a few
seconds:

a) Parsoid stops accepting new requests while restarting. This typically
takes a few seconds, but can take up to 60 seconds if all workers are busy.

b) After the restart the first request per worker will be slightly
slower as the JIT is doing its thing.

With Varnish backend retries some of this might not result in
client-visible failures, although I am not sure that Varnish retries
non-idempotent POSTs. That would mean that page saves are failing.

I think I'll rather (ask a root to) do a

  dsh -g parsoid service parsoid restart

instead until the salt bug is fixed.

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

Reply via email to