Re: real world testing of Redis as a bayes backend?

[Jan Hejl]

Dne 13.2.2015 v 11:38 Axb napsal(a):

On 02/13/2015 11:12 AM, Jan Hejl wrote:

Dne 13.2.2015 v 09:47 Axb napsal(a):

On 02/13/2015 08:51 AM, Jan Hejl wrote:

Whups. 62M :-D Sorry, my mistake. But still ...

what counts is:

used_memory_human:49.70M


seems amazingly little data.

Why not?

no scientific reason - in my setup I'd be worried that something isn't right because spam for Bayes is mostly fed from dedicated trap traffic and normally 49 MB would fill up in a few minutes.

I did in past, but that could be caused by empting bayes_seen (when I was migrating from MySQL, tokens were kept), no autolearn and (almost) on error learning with short expire times (see my settings):

bayes_auto_learn                0
bayes_token_ttl                 60d
bayes_seen_ttl                  14d

Coud it be?

BAYES hits as expected, this is reason to not worry until you said I should :-D

Dne 13.2.2015 v 08:35 Axb napsal(a):

Jan,

could you please post your "# Memory"  section of

"redis-cli info"

and

sa-learn --dump magic


On 02/13/2015 08:29 AM, Jan Hejl wrote:

Hi Quannah,

we use it on heavy loaded infrstructure - 10M messsages per day and
the
RES is 62k VIRT 130M. There has to be something "wrong" with his
configuration.

Jan

Dne 12.2.2015 v 23:04 Quanah Gibson-Mount napsal(a):

Has anyone done any real testing of Redis as a bayes backend? Talking

with one of our customer, with a trivial <60,000 accounts, they are
seeing:

PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM TIME+ COMMAND
22452 redis     20   0 28.7g  28g  740 S  9.6 72.4 1139:38
redis-server

28GB purely for a bayes DB for < 60k accounts is insanity.

--Quanah

--

Quanah Gibson-Mount
Platform Architect
Zimbra, Inc.
--------------------
Zimbra ::  the leader in open source messaging and collaboration

smime.p7s
Description: Elektronicky podpis S/MIME