Re: [basedb-devel] JobAgents on multi-processor machine

2007-10-31 Thread Paweł Sztromwasser
2007/10/29, Nicklas Nordborg <[EMAIL PROTECTED]>: > Paweł Sztromwasser wrote: > > Hi, > > > > I am trying to test the performance of Base2 (2.4.3) installation on > > our server. It is an 8-processor machine so we thought about running > > several job agents on it to speed up the plugin execution.

Re: [basedb-devel] JobAgents on multi-processor machine

2007-10-31 Thread Nicklas Nordborg
Paweł Sztromwasser wrote: > All job agents were configured to run all installed plugins and > Everyone group has access to them. In fact each of them is able to run > a job, but only if no other job is running. I saw couple of times (on > JobAgents tab in web UI), as 4 of jobs were assigned to (le

Re: [basedb-devel] JobAgents on multi-processor machine

2007-10-31 Thread Paweł Sztromwasser
When you run one job at a time it takes about 40-45 sec. Here are the times for 5 users trying to import simultanously no created started ended ended - started(sec) 1 11:59:16 11:59:22 12:00:02 40 2 11:59:16 11:59:30 12:00:40 70 3 11:59:16 11:59:45 12:01:17 92 4 11:59:16 11:59:42 12

[basedb-devel] BASE 2.4.5 available for download

2007-10-31 Thread Jari Häkkinen
A new version of BASE is available for download at http://base.thep.lu.se/wiki/DownloadPage This release contains one fix removing a debug sleep statement. Server performance is improved when upgrading to this release. The Lund BASE team

Re: [basedb-devel] JobAgents on multi-processor machine

2007-10-31 Thread Nicklas Nordborg
Paweł Sztromwasser wrote: > When you run one job at a time it takes about 40-45 sec. Here are the > times for 5 users trying to import simultanously > > no created started ended ended - started(sec) > 1 11:59:16 11:59:22 12:00:02 40 > 2 11:59:16 11:59:30 12:00:40 70 > 3 11:59:16 11:59:4

Re: [basedb-devel] JobAgents on multi-processor machine - private:)

2007-10-31 Thread Nicklas Nordborg
Pawel Sztromwasser wrote: > Nicklas Nordborg wrote: >> Pawel Sztromwasser wrote: >>> Hey, >>> >>> I came private to not spam others. Could you send me source code of >>> tests you used? I would like to check if this is not me doing >>> something wrong with the API before I get deep into postgreSQ