Re: Listen / Notify rewrite - Mailing list pgsql-hackers

From Greg Sabino Mullane
Subject Re: Listen / Notify rewrite
Date
Msg-id c96ab72010e16e08ad8b4ab37cc141e8@biglumber.com
Whole thread Raw
In response to Re: Listen / Notify rewrite  (Merlin Moncure <mmoncure@gmail.com>)
List pgsql-hackers
-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160


> old method scaled (badly) on volume of notifications and your stuff
> seems to scale based on # of client's sending simultaneous
> notifications.   Well, you're better all day long, but it shows that
> your fears regarding locking were not completely unfounded.  Do the
> Burcardo people have any insights on the #of simultaneous notifies are
> generated from different backends?

Very low. On a busy system I know of there are about 90 entries in
pg_listener, and I would guess that the maximum rate of simulataneous
notifies is not more than 3 per second, tops. If someone knows an
easy way to measure such a thing and is really curious, I can see about
getting better numbers.

- --
Greg Sabino Mullane greg@turnstep.com
End Point Corporation
PGP Key: 0x14964AC8 200911162127
http://biglumber.com/x/web?pk=2529DF6AB8F79407E94445B4BC9B906714964AC8
-----BEGIN PGP SIGNATURE-----

iEYEAREDAAYFAksCCjsACgkQvJuQZxSWSsgTogCfS5Xg8N2JhsUpi2r96IbxX+Tm
pMsAnAktBVkEblzx6Ux/netXkP9u4AVG
=SO/k
-----END PGP SIGNATURE-----




pgsql-hackers by date:

Previous
From: Greg Smith
Date:
Subject: Re: Partitioning option for COPY
Next
From: Itagaki Takahiro
Date:
Subject: Re: UTF8 with BOM support in psql