Re: pg_dump pg_restore hanging in CentOS for large data

Hi Tom,
Please find my answers below.

->ulimit : Verified in file /etc/security/limits.conf - it has been
configured as 20000. Should I increase the limit and try again? please let
me know.
root - nofile 20000
postgres - nofile 20000

->ssl_renegotiation_limit : Its been set as 512MB, but it has
been commented out in postgres.conf.
We are using CiscoSSL in our product.

# - Security and Authentication -

#authentication_timeout = 1min # 1s-600s
#ssl = off # (change requires restart)
#ssl_ciphers = 'ALL:!' # allowed SSL ciphers
# (change requires restart)
#ssl_renegotiation_limit = 512MB

Thanks,
Sridevi

On Thu, Mar 16, 2017 at 3:18 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:

> Sridevi B <sridevi17(at)gmail(dot)com> writes:
> > I am facing an issue with backup/Restore for data size more than
> *2GB*.
>
> I'm suspicious that you're running into some limit external to postgres
> itself. A couple of likely possibilities:
>
> * maybe you've got a ulimit active on the server or the pg_dump or
> pg_restore proces.
>
> * if you're using SSL encryption on the connection, you might be running
> into known bugs in openssl's management of renegotiation. This is made
> more likely by the fact that you're running such an old OS release (and
> PG 9.2.9 is not exactly up to date either). We changed the default
> value of ssl_renegotiation_limit to zero around 9.2.14, and you might
> want to set that explicitly in your 9.2.9 server.
>
> regards, tom lane
>