Thread: BUG #5853: when the walsender cannot exit when reboot slave computer

The following bug has been logged online:

Bug reference:      5853
Logged by:          lxzou
Email address:      zoulx1982@163.com
PostgreSQL version: 9.0.1
Operating system:   Linux 64
Description:        when the walsender cannot exit when reboot slave
computer
Details:

In streaming replication, when i restart the computer where slave in
running, the process walsender in primary cannot exit.

On 27.01.2011 15:18, lxzou wrote:
> The following bug has been logged online:
>
> Bug reference:      5853
> Logged by:          lxzou
> Email address:      zoulx1982@163.com
> PostgreSQL version: 9.0.1
> Operating system:   Linux 64
> Description:        when the walsender cannot exit when reboot slave
> computer
> Details:
>
> In streaming replication, when i restart the computer where slave in
> running, the process walsender in primary cannot exit.

How exactly does the problem manifest itself? Do you get an error
message? Anything else in the logs?

--
   Heikki Linnakangas
   EnterpriseDB   http://www.enterprisedb.com

(Please keep the mailing list cc'd so that others can help)

On 27.01.2011 16:38, zoulx1982 wrote:
> There is no error or log, but when  i use  "ps -ef | grep postgres" to watch the process, i find walsender process
alreadyexists. 
> In my view, whether this process should exit, because the walreceiver have exit  (caused by computer restart).

If the client just goes away, like if the computer is reset, it can take
up to a few minutes for the operating system to declare the TCP
connection as dead. And you'll need to generate a little bit of WAL so
that the walsender has something it tries to send to the receiver,
otherwise it might take even longer. You can also look at "netstat"
output, to see if the OS thinks the connection is still established.

--
   Heikki Linnakangas
   EnterpriseDB   http://www.enterprisedb.com