[rt-users] Performance on PostgreSQL

[Vivek Khera]

On Oct 13, 2006, at 12:18 AM, Jesse Vincent wrote:

>
> Strongly recommend:
> 	* newer RT
> 	* newer Pg (8.x before 8.1.4 had some specialness)
> 	* vacuum analyzing your postgres instance.

Also make sure your Pg is tuned properly for the size of your DB.   
See recent discussion in the postgres general mailing list (some tips  
came flying by earlier this week.  Also see http://www.varlena.com/ 
GeneralBits/Tidbits/annotated_conf_e.html.

I personally bump up the shared buffers, work mem, sort mem,  
checkpoint segments, and appropriatly set the effective cache size.

If you're still running slow after bumping these up (and applying my  
indexes, mentioned earlier), you should profile your overall system  
to see where the bottleneck is.  I'll vote it is your disk drives.   
At least use SATA...

Our primary RT install is running on a Dell PE 1850 with hardware  
RAID1 on a pair of SCSI disks which can support 80Mb/s sustained read/ 
write speed. With several tens of thousand tickets, it is quite  
snappy.  We even run it over SSL.

Our configuration of software is FreeBSD 6.1 + apache + fastcgi for  
RT.  RT makes zero use of mod_perl's advanced features, and fastcgi  
is much easier to get working right.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 2530 bytes
Desc: not available
URL: <http://lists.bestpractical.com/pipermail/rt-users/attachments/20061013/9bdbc19c/attachment.bin>