SEO

Erm, no, it is a Solaris box.
And as I could experience, there are rather long timeouts in solaris' tcp
implementation.

Thanks anyway.

What about Postgres. Wouldn't it be better to have postgres sending keep
alive or heart beat packages during that time?

|-----Original Message-----
|From: Dario Brignardello [mailto:dbrignar@uolsinectis.com]
|Sent: Mittwoch, 03. August 2005 19:33
|To: KÖPFERL Robert
|Cc: pgsql-admin@postgresql.org
|Subject: Re: [ADMIN] Blocking connection and timeout problem
|
|
|Assuming you are on a linux box (Not sure if it's stated in your mail,
|sorry if I miss that) you could set
|
|/proc/sys/net/ipv4/tcp_keepalive_time
|
|to a lower value. That should do the trick :-)
|
|
|Hope it helps.
|
|Greetings
|Dario.
|
|
|
|On Wed, 2005-08-03 at 16:48 +0200, KÖPFERL Robert wrote:
|>
|>
|>
|> |-----Original Message-----
|> |From: Tom Lane [mailto:tgl@sss.pgh.pa.us]
|> |Sent: Mittwoch, 03. August 2005 16:38
|> |To: KÖPFERL Robert
|> |Cc: pgsql-admin@postgresql.org
|> |Subject: Re: [ADMIN] Blocking connection and timeout problem
|> |
|> |
|> |=?iso-8859-1?Q?K=D6PFERL_Robert?=
|<robert.koepferl@sonorys.at> writes:
|> |> Why is that? Why isn't client A's transaction detected as dead ?
|> |
|> |It will be eventually, when the TCP connection times out.
|The standard
|> |timeout is generally an hour or two :-(
|>
|> Seems not like so. I waited for two hours and it still existed.
|>
|> But non-the-less. Can this timeout be configured to be less?
|>
|> |
|> | regards, tom lane
|> |
|>
|> ---------------------------(end of
|broadcast)---------------------------
|> TIP 6: explain analyze is your friend
|>
|>
|--
|Atte:
|Dario Brignardello
|Planificacion y Desarrollo
|Tecnologia
|UOL Argentina S.A
|
|Florida 537 Piso 6, Buenos Aires, Argentina
|+54-11-4321-9110 ext 2533
|PGP public key: http://webs.uolsinectis.com.ar/dbrignar/pgp.html
|

---------------------------(end of broadcast)---------------------------
TIP 4: Have you searched our list archives?

http://archives.postgresql.org

Taking the risk of running offtopic: That value is defined in Solaris as
well. Use:

ndd -get /dev/tcp tcp_keepalive_interval

to retrieve it, and

/usr/sbin/ndd -set /dev/tcp tcp_keepalive_interval 3600000

to set it to one hour. Check
http://sunsolve.sun.com/search/docum...ey=1-30-2876-1
for more information.

Regarding the possibility of heartbeats in postgres ... have no idea how
wise would it be ... there's people more qualified here that could
answer that ;-D

Hope this helps, anyway.
Regards.
Dario





On Fri, 2005-08-05 at 09:43 +0200, KÖPFERL Robert wrote:
>
> Erm, no, it is a Solaris box.
> And as I could experience, there are rather long timeouts in solaris' tcp
> implementation.
>
> Thanks anyway.
>
> What about Postgres. Wouldn't it be better to have postgres sending keep
> alive or heart beat packages during that time?
>
> |-----Original Message-----
> |From: Dario Brignardello [mailto:dbrignar@uolsinectis.com]
> |Sent: Mittwoch, 03. August 2005 19:33
> |To: KÖPFERL Robert
> |Cc: pgsql-admin@postgresql.org
> |Subject: Re: [ADMIN] Blocking connection and timeout problem
> |
> |
> |Assuming you are on a linux box (Not sure if it's stated in your mail,
> |sorry if I miss that) you could set
> |
> |/proc/sys/net/ipv4/tcp_keepalive_time
> |
> |to a lower value. That should do the trick :-)
> |
> |
> |Hope it helps.
> |
> |Greetings
> |Dario.
> |
> |
> |
> |On Wed, 2005-08-03 at 16:48 +0200, KÖPFERL Robert wrote:
> |>
> |>
> |>
> |> |-----Original Message-----
> |> |From: Tom Lane [mailto:tgl@sss.pgh.pa.us]
> |> |Sent: Mittwoch, 03. August 2005 16:38
> |> |To: KÖPFERL Robert
> |> |Cc: pgsql-admin@postgresql.org
> |> |Subject: Re: [ADMIN] Blocking connection and timeout problem
> |> |
> |> |
> |> |=?iso-8859-1?Q?K=D6PFERL_Robert?=
> |<robert.koepferl@sonorys.at> writes:
> |> |> Why is that? Why isn't client A's transaction detected as dead ?
> |> |
> |> |It will be eventually, when the TCP connection times out.
> |The standard
> |> |timeout is generally an hour or two :-(
> |>
> |> Seems not like so. I waited for two hours and it still existed.
> |>
> |> But non-the-less. Can this timeout be configured to be less?
> |>
> |> |
> |> | regards, tom lane
> |> |
> |>
> |> ---------------------------(end of
> |broadcast)---------------------------
> |> TIP 6: explain analyze is your friend
> |>
> |>
> |--
> |Atte:
> |Dario Brignardello
> |Planificacion y Desarrollo
> |Tecnologia
> |UOL Argentina S.A
> |
> |Florida 537 Piso 6, Buenos Aires, Argentina
> |+54-11-4321-9110 ext 2533
> |PGP public key: http://webs.uolsinectis.com.ar/dbrignar/pgp.html
> |
>
> ---------------------------(end of broadcast)---------------------------
> TIP 4: Have you searched our list archives?
>
> http://archives.postgresql.org
>



---------------------------(end of broadcast)---------------------------
TIP 6: explain analyze is your friend