Parallel pg_basebackup

Started by Emanuel Calvoover 12 years ago3 messages
#1Emanuel Calvo
postgres.arg@gmail.com

Hi guys,

I was wondering if there is a proposal for parallelize pg_basebackup (sorry
if there is already one, I looked into the inbox and did not find any).

I made a simple test and the performance gain is pretty considerable using
parallel rsync processes vs. single pg_basebackup.

i.e., for 46GB of data I had the following results:

Using https://gist.github.com/rcoup/5358786:
rsync, parallel, including the xlogs
real 18m52.674s
user 33m54.728s
sys 13m44.604s

(-x option)
real 23m26.802s
user 0m24.392s
sys 3m55.116s

Actually, AWS services limits the bandwidth per process but if you have
parallel processes, none of them will have the IOPS decreased which will be
translated in a great improvement. Also the network bandwidth is strong
between the regions.

Thanks guys!

--
--
Emanuel Calvo

#2Peter Eisentraut
peter_e@gmx.net
In reply to: Emanuel Calvo (#1)
Re: Parallel pg_basebackup

On Fri, 2013-08-23 at 10:51 -0300, Emanuel Calvo wrote:

I was wondering if there is a proposal for parallelize pg_basebackup

There isn't one, but after some talk behind the scenes, I think we
should definitely look into it.

--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

#3Emanuel Calvo
postgres.arg@gmail.com
In reply to: Peter Eisentraut (#2)
Re: Parallel pg_basebackup

That looks promising! Thanks Peter!

2013/8/24 Peter Eisentraut <peter_e@gmx.net>

On Fri, 2013-08-23 at 10:51 -0300, Emanuel Calvo wrote:

I was wondering if there is a proposal for parallelize pg_basebackup

There isn't one, but after some talk behind the scenes, I think we
should definitely look into it.

--
--
Emanuel Calvo