From: | Adrian Klaver <adrian(dot)klaver(at)aklaver(dot)com> |
---|---|
To: | Nicola Contu <nicola(dot)contu(at)gmail(dot)com> |
Cc: | pgsql-general <pgsql-general(at)postgresql(dot)org>, Alessandro Aste <Alessandro(dot)aste(at)gtt(dot)net> |
Subject: | Re: Streaming replication - 11.5 |
Date: | 2020-03-16 15:28:43 |
Message-ID: | bced3a6e-0c25-e665-f666-52126c46d828@aklaver.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On 3/16/20 2:57 AM, Nicola Contu wrote:
> I was able to make pg_basebackup working using --max-rate=128M
>
> Still don't understand why. I guess it is related to the encryption and
> slowness of the disk..
> Do you have any idea?
I think your explanation fits. Encryption/decryption have overhead.
>
> Il giorno ven 13 mar 2020 alle ore 16:15 Adrian Klaver
> <adrian(dot)klaver(at)aklaver(dot)com <mailto:adrian(dot)klaver(at)aklaver(dot)com>> ha scritto:
>
> On 3/13/20 4:11 AM, Nicola Contu wrote:
> > So in the logs I now see this :
> >
> > 2020-03-13 11:03:42 GMT [10.150.20.22(45294)] [27804]: [1-1]
> > db=[unknown],user=replicator LOG: terminating walsender process
> due to
> > replication timeout
>
> Yeah that's been showing up the log snippets you have been posting.
>
> To figure this out you will need to:
>
> 1) Make a list of what changed since the last time replication worked
> consistently.
>
> 2) Monitor the changed components, start logging or increase logging.
>
> 3) Monitor the chain of replication as whole, to catch changes that you
> do not know about. Since you seem to be operating across data centers
> that would include verifying the network.
>
> >
> > So I tried increasing the wal_sender_timeout to 300s but it did
> not help
> >
> > Il giorno gio 12 mar 2020 alle ore 15:56 Nicola Contu
> > <nicola(dot)contu(at)gmail(dot)com <mailto:nicola(dot)contu(at)gmail(dot)com>
> <mailto:nicola(dot)contu(at)gmail(dot)com <mailto:nicola(dot)contu(at)gmail(dot)com>>> ha
> scritto:
> >
> > The encryption is at os level. So the drives are encrypted with a
> > password where the db saves data
> >
> > Il gio 12 mar 2020, 15:51 Adrian Klaver
> <adrian(dot)klaver(at)aklaver(dot)com <mailto:adrian(dot)klaver(at)aklaver(dot)com>
> > <mailto:adrian(dot)klaver(at)aklaver(dot)com
> <mailto:adrian(dot)klaver(at)aklaver(dot)com>>> ha scritto:
> >
> > On 3/12/20 4:31 AM, Nicola Contu wrote:
> > > The replicator is ok and the replicated as well.
> > > %Cpu(s): 0.2 us, 1.0 sy, 0.0 ni, 94.8 id, 4.0 wa, 0.0
> > hi, 0.0 si,
> > > 0.0 st
> > >
> > > CPU is really low on both.
> > >
> > > I am running pg_basebackup again everytime.
> > > Any other suggestions?
> > >
> >
> > I have to believe their is a connection between changing to
> > encrypting
> > the disks and your issues. Not sure what, but to help how
> is the
> > encryption being done and what program is being used?
> >
> >
> > --
> > Adrian Klaver
> > adrian(dot)klaver(at)aklaver(dot)com <mailto:adrian(dot)klaver(at)aklaver(dot)com>
> <mailto:adrian(dot)klaver(at)aklaver(dot)com <mailto:adrian(dot)klaver(at)aklaver(dot)com>>
> >
>
>
> --
> Adrian Klaver
> adrian(dot)klaver(at)aklaver(dot)com <mailto:adrian(dot)klaver(at)aklaver(dot)com>
>
--
Adrian Klaver
adrian(dot)klaver(at)aklaver(dot)com
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2020-03-16 15:28:54 | Re: Mixed Locales and Upgrading |
Previous Message | Adrian Klaver | 2020-03-16 15:27:04 | Re: Order by and timestamp |