Hi Daniel,
I already tried to set the destination table to unlogged - it improved the
performance slightly. Is there a way to make sure that I/O is the problem ?
2017-08-17 0:46 GMT+03:00 Daniel Blanch Bataller <
daniel.blanch.batal...@gmail.com>:
> Seems your disks are too slow. On my laptop (no
Seems your disks are too slow. On my laptop (nothing special, just one disk)
using COPY I can dump 3G in ~ 20 secs, loading takes 120 secs, bare copying 3G
takes 10 secs.
Similar proportion you had, but much faster.
confirm I/O is your bottleneck, and tell us how you solved your problem
Anyw
My server is virtual and it have virtual hd from a vnx storage machine. The
logs and the data are on the same disk.
2017-08-16 17:04 GMT+03:00 Daniel Blanch Bataller <
daniel.blanch.batal...@gmail.com>:
> Considering it has to write logs and data at checkpoints I don’t see it
> particularly slow
Considering it has to write logs and data at checkpoints I don’t see it
particularly slow compared to the extract phase. What kind of disks you have
SSD or regular disks? Different disks for ltransaction logs and data?
> El 16 ago 2017, a las 15:54, Mariel Cherkassky
> escribió:
>
> I run th
I run the copy command via psql to create a local dump of a 3G table and it
took me 134059.732ms =~2 minutes. After that I imported the data via copy
and it took 458648.677ms =~7 minutes. So the copy command works but pretty
slow.
2017-08-16 16:08 GMT+03:00 Daniel Blanch Bataller <
daniel.blanch.b
See if the copy command is actually working, copy should be very fast from your
local disk.
> El 16 ago 2017, a las 14:26, Mariel Cherkassky
> escribió:
>
>
> After all the changes of the memory parameters the same operation(without the
> copy utility) didnt run much faster - it took one m
After all the changes of the memory parameters the same operation(without
the copy utility) didnt run much faster - it took one minute less. I made
a test with the copy command (without the 'with binary') and it took 1.5
hours to create the dumpfile in my local postgresql server. Then I tried to
r
