Hello again... Quoting Andy Allan <[email protected]>:
How can I investigate what the bottleneck is? So far, I can see, that there is a lot of I/O going on and not much CPU usage. The diff import is stuck upon prompting "Going over pending relations","processing relation (...)".
In my experience it's all about the disks. What disks do you have, speed of them etc? I found that 7.2k SATA disks were just about not quite fast enough, whereas 10k SATA, SAS or SSDs seem to be fine.
Ok, so I finally set up a (software) RAID-0 with three disks. Nothing too fancy, but still, I am not completely satisifed. I now can barely keep up with the update rate, but Postgres doesn't seem to utilize the RAID completely (CPU load is next to 0% and RAM usage is also not exhausted). Looking at iostat and device Utilization, all drives are at about 30-35% (compared to 100% with the single drive setup. But when writing a file to the raid with dd I get 100%.
Do I have to tell Postgres, that there are more discs to do optimizations or do I have to reimport, to rearrange the data in a smarter way?
Any pointers are highly appreciated :-) Markus _______________________________________________ dev mailing list [email protected] http://lists.openstreetmap.org/listinfo/dev
