Hi John,
How frequently are you writing the stats to the DB ? If you think it is purely
SQL performance related then keep reading.
What are your "sql_history" & "sql_history_roundoff" values ?
You can also use "sql_dont_try_update" to force it to use an "insert" statement
instead of "update".
The setting "sql_refresh_time" controls how often data is purged to the DB.
We run a fairly small environment. The throughput we are monitoring is about
10% of yours and using hourly rounded stats aggregated by destination we get
about 200M per month of DB usage. It runs on an older server with a single 2GHz
CPU & 2G of RAM. Disk is (I think) 2x 146G SCSI drives (RAID-1).
What are you trying to monitor and record ? What sort of granularity do you
need in your data ?
Others on the list might have some other suggestions ? Possibly posting your
config file (less anything that is sensitive) might help people to point out
what you might change ?
regards,
Tony.
>________________________________
> From: John Center <[email protected]>
>To: [email protected]
>Sent: Wednesday, 10 October 2012 2:37 AM
>Subject: Re: [pmacct-discussion] Sizing pmacct server
>
>Hi Tony,
>
>The aggregation is: tag, src_host, dst_host, proto, dst_port
>
>The database has grown to 8.4G in 10 days. I'm not sure how quickly we
>are adding to the DB. The 6500 i/o is around 400-600Mbs.
>
>Thanks.
>
> -John
>
>
>On 10/04/2012 08:23 PM, Tony wrote:
>> Hi John,
>>
>> How are you aggregating/summarising your data ?
>>
>> What is the traffic throughput on your 6500 ? What is the volume of
>> data you are seeing ? How big/quickly is your DB growing ?
>>
>>
>> regards,
>> Tony.
>>
>>
>> ------------------------------------------------------------------------
>> *From:* John Center <[email protected]>
>> *To:* [email protected]
>> *Sent:* Friday, 5 October 2012 4:48 AM
>> *Subject:* [pmacct-discussion] Sizing pmacct server
>>
>> Hi,
>>
>> We were wondering what size system people are using for pmacct?
>> We've been testing pmacct with mysql & pmgraph on a Dell PowerEdge
>> 1900, but can only keep pmacct running for an hour or so before
>> restarting nfacctd. It appears that mysql can't keep up with
>> nfacctd, with the number of rows getting too large too quickly. We
>> are only receiving flows from a single Cat 6506.
>>
>> Thanks.
>>
>> -John
>>
>> -- John Center
>> Villanova University
>>
>> _______________________________________________
>> pmacct-discussion mailing list
>> http://www.pmacct.net/#mailinglists
>>
>>
>>
>>
>> _______________________________________________
>> pmacct-discussion mailing list
>> http://www.pmacct.net/#mailinglists
>>
>
>_______________________________________________
>pmacct-discussion mailing list
>http://www.pmacct.net/#mailinglists
>
>
>
_______________________________________________
pmacct-discussion mailing list
http://www.pmacct.net/#mailinglists
