Re: [SOLVED] pgAdmin ver 1.10 freezing all too frequently

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Michael,

My postgresql.conf, tcp-related, bit:
tcp_keepalives_idle = 75 # TCP_KEEPIDLE, in seconds;
# 0 selects the system default
tcp_keepalives_interval = 60 # TCP_KEEPINTVL, in seconds;
# 0 selects the system default
#tcp_keepalives_count = 0 # TCP_KEEPCNT;
# 0 selects the system default

BR,
Pedro.

On 11/12/2009 01:08 PM, Michael Shapiro wrote:
> My server's setting are:
>
> 'tcp_keepalives_count','9' 'tcp_keepalives_idle','7200'
> 'tcp_keepalives_interval','75'
>
> and I still get the problem. I just have to wait 20 minutes or so.
>
> (I don't really know how to predict how long it takes to lose a
> connection, even knowing these settings)
>
> Note that if someone wants to replicate my problem, they could use
> these settings and run PgAdmin from a Windows box (since the
> settings don't mean anything for Unix connections).
>
> Pedro, what are your setttings?
>
> On Thu, Nov 12, 2009 at 5:40 AM, Pedro Doria Meunier
> <pdoria(at)netmadeira(dot)com <mailto:pdoria(at)netmadeira(dot)com>> wrote:
>
> Hi Dave,
>
> I was thinking about Michael's observations and the fact that this
> problem never happened if the server status window was opened.
>
> So what I did was: set the "tcp_keepalives_interval" = 75 and
> reloaded the server's configuration.
>
> No more freezes ever since! :)
>
> IMHO, however, pgAdmin should better handle connection status or,
> alternatively, check it and initiate a new connection if the
> previous one stalled...
>
> I have *some* C proficiency so if you point me in the right
> direction I might start investigating it and try to come up with a
> solution ... ;)
>
> BR, Pedro.
>
> On 11/12/2009 11:17 AM, Dave Page wrote:
>> On Thu, Nov 12, 2009 at 11:06 AM, Pedro Doria Meunier
>> <pdoria(at)netmadeira(dot)com <mailto:pdoria(at)netmadeira(dot)com>> wrote:
>>> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
>>>
>>> I'm sorry Dave.
>>>
>>> I'm unable to... :(
>>>
>>> Whenever I try to get a backtrace it's already frozen solid
>>> and there's nothing to do other than kill it...
>
>> If you start it under gdb, doesn't gdb break when you kill
>> pgAdmin? Something like:
>
>> Program received signal SIGKILL, Killed. 0x943407da in
>> mach_msg_trap () (gdb) (gdb) bt #0 0x943407da in mach_msg_trap
>> () #1 0x94340f47 in mach_msg () #2 0x99006dbf in __CFRunLoopRun
>> () #3 0x99005d34 in CFRunLoopRunSpecific () #4 0x99005b61 in
>> CFRunLoopRunInMode () #5 0x919f3fec in
>> RunCurrentEventLoopInMode () #6 0x919f3da3 in
>> ReceiveNextEventCommon () #7 0x91b7bd91 in ReceiveNextEvent ()
>> #8 0x00e9d7e7 in wxApp::MacDoOneEvent () #9 0x00eb6fd3 in
>> wxEventLoop::Dispatch () #10 0x00f5d708 in wxEventLoopManual::Run
>> () #11 0x00f36bf3 in wxAppBase::MainLoop () #12 0x0136981a in
>> wxEntry () #13 0x0003f548 in main ()
>

- --
Sent via pgadmin-support mailing list
(pgadmin-support(at)postgresql(dot)org
<mailto:pgadmin-support(at)postgresql(dot)org>)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgadmin-support

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org/

iEYEARECAAYFAkr8DB0ACgkQ2FH5GXCfxAsNugCgqfajYwo0fuhw/BmH2elF9JlA
0zUAnA13I5GDbOytVkK+az2P/UcatjWs
=80sf
-----END PGP SIGNATURE-----