[pgpool-hackers: 1573] Re: Pgpool Hangs connections when shutting down interface of the secondary
Tatsuo Ishii
ishii at postgresql.org
Fri May 20 12:15:06 JST 2016
> I did the test with watchdog enabled and disabled. The problem is, with
> watchdog disabled, there is no virtual IP, so, the test is pointless.
My point is not that you should operate without watchdog, rather than
for identifying problems purpose. If the problem does not occur
without watchdog, watchdog could be responsible for the problem.
Best regards,
--
Tatsuo Ishii
SRA OSS, Inc. Japan
English: http://www.sraoss.co.jp/index_en.php
Japanese:http://www.sraoss.co.jp
> I
> have a bug ticket open, with config and everything.
You seem not uploaded pgpool log.
Best regards,
--
Tatsuo Ishii
SRA OSS, Inc. Japan
English: http://www.sraoss.co.jp/index_en.php
Japanese:http://www.sraoss.co.jp
> Regards
> On May 19, 2016 11:31 PM, "Tatsuo Ishii" <ishii at postgresql.org> wrote:
>
>> Do you enable watchdog? If yes, can you do the same test with watchdog
>> disabled and see what happens?
>>
>> Best regards,
>> --
>> Tatsuo Ishii
>> SRA OSS, Inc. Japan
>> English: http://www.sraoss.co.jp/index_en.php
>> Japanese:http://www.sraoss.co.jp
>>
>> > Of course not. I shut down the interface. But the point of it is, I would
>> > expect pgpool to mark node 2 as down, and that is not happening.
>> > Thanks for your answers Tatsuo.
>> > Regards!
>> > On May 19, 2016 11:11 PM, "Tatsuo Ishii" <ishii at postgresql.org> wrote:
>> >
>> >> So if you run psql on the node 1, can it connect to PostgreSQL (not
>> >> via pgpool) on node 2?
>> >>
>> >> Best regards,
>> >> --
>> >> Tatsuo Ishii
>> >> SRA OSS, Inc. Japan
>> >> English: http://www.sraoss.co.jp/index_en.php
>> >> Japanese:http://www.sraoss.co.jp
>> >>
>> >> > Hey Tatsuo
>> >> > The initial situation is that node 1 is master node for pgpool and
>> >> primary
>> >> > node for postgres. Node 2 is standby node for pgpool and slave for
>> >> > postgres. If I shut down the interface on the slave/standby mode, next
>> >> time
>> >> > I try to connect to postgres through pgpool, the connections just
>> hang. I
>> >> > would think this would be pretty repeatable, but it looks like no one
>> >> else
>> >> > is having this issue?
>> >> > Thanks.
>> >> > Regards
>> >> > On May 19, 2016 10:36 PM, "Tatsuo Ishii" <ishii at postgresql.org>
>> wrote:
>> >> >
>> >> >> I just want to make sure. I assume psql is running on the same
>> machine
>> >> >> as pgpool is running on. Am I correct? I am asking because it sounds
>> >> >> ridiculous that psql can connect to the PostgreSQL server which its
>> >> >> network interface is down.
>> >> >>
>> >> >> Best regards,
>> >> >> --
>> >> >> Tatsuo Ishii
>> >> >> SRA OSS, Inc. Japan
>> >> >> English: http://www.sraoss.co.jp/index_en.php
>> >> >> Japanese:http://www.sraoss.co.jp
>> >> >>
>> >> >> > Hey Tatsuo, thanks for the feedback, I really appreciate it.
>> >> Connecting
>> >> >> > directly, it works.
>> >> >> > Thanks! Regards
>> >> >> > On May 19, 2016 9:24 PM, "Tatsuo Ishii" <ishii at postgresql.org>
>> wrote:
>> >> >> >
>> >> >> >> If you directly connect to PostgreSQL on the server on which the
>> >> >> >> network interface was down using psql, does it fail or hang?
>> >> >> >>
>> >> >> >> Best regards,
>> >> >> >> --
>> >> >> >> Tatsuo Ishii
>> >> >> >> SRA OSS, Inc. Japan
>> >> >> >> English: http://www.sraoss.co.jp/index_en.php
>> >> >> >> Japanese:http://www.sraoss.co.jp
>> >> >> >>
>> >> >> >> > Hello Guys:
>> >> >> >> > i have posted this issue in pgpool-Users, but i got no answer.
>> At
>> >> this
>> >> >> >> > point i am thinking about switching solutions as i cannot make
>> mine
>> >> >> work,
>> >> >> >> > but i really like the idea of using pgpool, so i was wondering
>> if i
>> >> >> could
>> >> >> >> > get an answer here. This is the issue i am experiencing:
>> >> >> >> >
>> >> >> >> > I am running pgpool 3.5.2 in a 2 node cluster with
>> postgresql-9.5
>> >> in
>> >> >> >> > master-slave and streaming replication mode.
>> >> >> >> >
>> >> >> >> > I have been testing failover and failback for a while. While
>> doing
>> >> >> >> failover
>> >> >> >> > by shutting down the processes, everything looks fine.
>> >> >> >> >
>> >> >> >> >
>> >> >> >> > The one test that is failing is when i shutdown the interface of
>> >> one
>> >> >> >> node.
>> >> >> >> > Right now i am shutting down the interface of the slave node
>> (both
>> >> >> pgool
>> >> >> >> > slave and postgresql slave).
>> >> >> >> >
>> >> >> >> > The problem i am running into is that after doing this, all
>> >> >> connections
>> >> >> >> to
>> >> >> >> > my database (through pgpool) hang. i am testing it with psql,
>> and
>> >> psql
>> >> >> >> just
>> >> >> >> > hangs and does not give me an output. When i bring the
>> interface of
>> >> >> the
>> >> >> >> > slave back up, and try to connect to psql again, it looks like
>> >> pgpool
>> >> >> >> never
>> >> >> >> > marked the postgresql node as disconnected.
>> >> >> >> >
>> >> >> >> > I tried both with and without health check, and also tried with
>> >> >> different
>> >> >> >> > values of health_check_timeout. my connect timeout value is
>> default
>> >> >> (10
>> >> >> >> > seconds).
>> >> >> >> >
>> >> >> >> > Any one has encountered this issue? I just dont see pgpool
>> >> attempting
>> >> >> to
>> >> >> >> do
>> >> >> >> > any failover. Pgpool is still running though, i can see the logs
>> >> still
>> >> >> >> > comming. I just never see an error.
>> >> >> >> >
>> >> >> >> > I am NOT using interface monitoring, and i would prefer not to
>> use
>> >> it.
>> >> >> >> >
>> >> >> >> > Any pointers on how could i troubleshoot this?
>> >> >> >>
>> >> >>
>> >>
>> >
>> > --001a114942f071445105333ca701
>> > Content-Type: text/html; charset=UTF-8
>> > Content-Transfer-Encoding: quoted-printable
>> >
>> > <p dir="ltr">Of course not. I shut down the interface. But the point of
>> it is, I would expect pgpool to mark node 2 as down, and that is not
>> happening.<br>
>> > Thanks for your answers Tatsuo.<br>
>> > Regards!</p>
>> > <div class="gmail_quote">On May 19, 2016 11:11 PM, "Tatsuo
>> Ishii" <<a href="mailto:ishii at postgresql.org">ishii at postgresql.org</a>>
>> wrote:<br type="attribution"><blockquote class="gmail_quote"
>> style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">So if
>> you run psql on the node 1, can it connect to PostgreSQL (not<br>
>> > via pgpool) on node 2?<br>
>> > <br>
>> > Best regards,<br>
>> > --<br>
>> > Tatsuo Ishii<br>
>> > SRA OSS, Inc. Japan<br>
>> > English: <a href="http://www.sraoss.co.jp/index_en.php"
>> rel="noreferrer" target="_blank">http://www.sraoss.co.jp/index_en.php
>> </a><br>
>> > Japanese:<a href="http://www.sraoss.co.jp" rel="noreferrer"
>> target="_blank">http://www.sraoss.co.jp</a><br>
>> > <br>
>> > > Hey Tatsuo<br>
>> > > The initial situation is that node 1 is master node for pgpool and
>> primary<br>
>> > > node for postgres. Node 2 is standby node for pgpool and slave
>> for<br>
>> > > postgres. If I shut down the interface on the slave/standby mode,
>> next time<br>
>> > > I try to connect to postgres through pgpool, the connections just
>> hang. I<br>
>> > > would think this would be pretty repeatable, but it looks like no
>> one else<br>
>> > > is having this issue?<br>
>> > > Thanks.<br>
>> > > Regards<br>
>> > > On May 19, 2016 10:36 PM, "Tatsuo Ishii" <<a
>> href="mailto:ishii at postgresql.org">ishii at postgresql.org</a>> wrote:<br>
>> > ><br>
>> > >> I just want to make sure. I assume psql is running on the same
>> machine<br>
>> > >> as pgpool is running on. Am I correct? I am asking because it
>> sounds<br>
>> > >> ridiculous that psql can connect to the PostgreSQL server which
>> its<br>
>> > >> network interface is down.<br>
>> > >><br>
>> > >> Best regards,<br>
>> > >> --<br>
>> > >> Tatsuo Ishii<br>
>> > >> SRA OSS, Inc. Japan<br>
>> > >> English: <a href="http://www.sraoss.co.jp/index_en.php"
>> rel="noreferrer" target="_blank">http://www.sraoss.co.jp/index_en.php
>> </a><br>
>> > >> Japanese:<a href="http://www.sraoss.co.jp" rel="noreferrer"
>> target="_blank">http://www.sraoss.co.jp</a><br>
>> > >><br>
>> > >> > Hey Tatsuo, thanks for the feedback, I really appreciate
>> it. Connecting<br>
>> > >> > directly, it works.<br>
>> > >> > Thanks! Regards<br>
>> > >> > On May 19, 2016 9:24 PM, "Tatsuo Ishii" <<a
>> href="mailto:ishii at postgresql.org">ishii at postgresql.org</a>> wrote:<br>
>> > >> ><br>
>> > >> >> If you directly connect to PostgreSQL on the server on
>> which the<br>
>> > >> >> network interface was down using psql, does it fail or
>> hang?<br>
>> > >> >><br>
>> > >> >> Best regards,<br>
>> > >> >> --<br>
>> > >> >> Tatsuo Ishii<br>
>> > >> >> SRA OSS, Inc. Japan<br>
>> > >> >> English: <a href="http://www.sraoss.co.jp/index_en.php"
>> rel="noreferrer" target="_blank">http://www.sraoss.co.jp/index_en.php
>> </a><br>
>> > >> >> Japanese:<a href="http://www.sraoss.co.jp"
>> rel="noreferrer" target="_blank">http://www.sraoss.co.jp</a><br>
>> > >> >><br>
>> > >> >> > Hello Guys:<br>
>> > >> >> > i have posted this issue in pgpool-Users, but i
>> got no answer. At this<br>
>> > >> >> > point i am thinking about switching solutions as
>> i cannot make mine<br>
>> > >> work,<br>
>> > >> >> > but i really like the idea of using pgpool, so i
>> was wondering if i<br>
>> > >> could<br>
>> > >> >> > get an answer here. This is the issue i am
>> experiencing:<br>
>> > >> >> ><br>
>> > >> >> > I am running pgpool 3.5.2 in a 2 node cluster
>> with postgresql-9.5 in<br>
>> > >> >> > master-slave and streaming replication mode.<br>
>> > >> >> ><br>
>> > >> >> > I have been testing failover and failback for a
>> while. While doing<br>
>> > >> >> failover<br>
>> > >> >> > by shutting down the processes, everything looks
>> fine.<br>
>> > >> >> ><br>
>> > >> >> ><br>
>> > >> >> > The one test that is failing is when i shutdown
>> the interface of one<br>
>> > >> >> node.<br>
>> > >> >> > Right now i am shutting down the interface of the
>> slave node (both<br>
>> > >> pgool<br>
>> > >> >> > slave and postgresql slave).<br>
>> > >> >> ><br>
>> > >> >> > The problem i am running into is that after doing
>> this, all<br>
>> > >> connections<br>
>> > >> >> to<br>
>> > >> >> > my database (through pgpool) hang. i am testing
>> it with psql, and psql<br>
>> > >> >> just<br>
>> > >> >> > hangs and does not give me an output. When i
>> bring the interface of<br>
>> > >> the<br>
>> > >> >> > slave back up, and try to connect to psql again,
>> it looks like pgpool<br>
>> > >> >> never<br>
>> > >> >> > marked the postgresql node as disconnected.<br>
>> > >> >> ><br>
>> > >> >> > I tried both with and without health check, and
>> also tried with<br>
>> > >> different<br>
>> > >> >> > values of health_check_timeout. my connect
>> timeout value is default<br>
>> > >> (10<br>
>> > >> >> > seconds).<br>
>> > >> >> ><br>
>> > >> >> > Any one has encountered this issue? I just dont
>> see pgpool attempting<br>
>> > >> to<br>
>> > >> >> do<br>
>> > >> >> > any failover. Pgpool is still running though, i
>> can see the logs still<br>
>> > >> >> > comming. I just never see an error.<br>
>> > >> >> ><br>
>> > >> >> > I am NOT using interface monitoring, and i would
>> prefer not to use it.<br>
>> > >> >> ><br>
>> > >> >> > Any pointers on how could i troubleshoot this?<br>
>> > >> >><br>
>> > >><br>
>> > </blockquote></div>
>> >
>> > --001a114942f071445105333ca701--
>>
More information about the pgpool-hackers
mailing list