Обсуждение: pgadmin3 crash report

Le 30/06/2010 10:16, r d a écrit :
> [...]
> pgadmin3 crashes when I try to see the server statistics (menu Tools -->
> Server Status).
> I think that I configured my installation correctly, details follow below.
> 
> My configuration:
> ---------------------------------------------------------------------------------
> - Linux FC12 installed clean, plus YUM upgrades from network
> uname -a == Linux L3 2.6.32.14-127.fc12.x86_64 #1 SMP Fri May 28
> 04:30:39 UTC 2010 x86_64 x86_64 x86_64 GNU/Linux
> 
> - PostgreSQL 8.4.4 installed via YUM from net
> postgresql-server-8.4.4-1.fc12.x86_64
> (and all postgresql-* packages plus dependencies)
> 
> - pgadmin3 1.10.3 - installed via YUM from net
> pgadmin3-1.10.3-2.fc12.x86_64
> 
> - wx 2.8.11 installed from net via YUM
> wxBase-2.8.11-1.fc12.x86_64
> wxPython-2.8.11.0-1.fc12.x86_64
> wxGTK-media-2.8.11-1.fc12.x86_64
> wxGTK-devel-2.8.11-1.fc12.x86_64
> wxGTK-gl-2.8.11-1.fc12.x86_64
> wxPython-devel-2.8.11.0-1.fc12.x86_64
> wxGTK-2.8.11-1.fc12.x86_64
> ---------------------------------------------------------------------------------
> 
> I set up the instrumentation, logging, ...  as per instructions, i.e.
> 1) psql --username=username dbname -f adminpack.sql   (actually, I have
> all contribs installed into all databases)
> 2) logging output filename is (postgresql.conf) log_filename =
> 'postgresql-%Y-%m-%d_%H%M%S.log'
> 3) logs are written into data/pg_log (default)
> 4) log files look like this (ls -l)
> 
> -rw-------. 1 postgres postgres       171 2010-06-30 01:41
> postgresql-2010-06-30_000000.log
> -rw-------. 1 postgres postgres    322532 2010-06-30 09:43
> postgresql-2010-06-30_084323.log
> 
> 5) Sample contents of a log content looks like this:
> STATEMENT:  SELECT
> pg_file_read('pg_log/postgresql-2010-06-30_084323.log', 0, 50000)
> LOG:  EXECUTOR STATISTICS
> DETAIL:  ! system usage stats:
>         !       0.000242 elapsed 0.000000 user 0.000000 system sec
>         !       [0.016997 user 0.005999 sys total]
>         !       0/0 [64/0] filesystem blocks in/out
>         !       0/4 [0/1563] page faults/reclaims, 0 [0] swaps
>         !       0 [0] signals rcvd, 0/0 [0/0] messages rcvd/sent
>         !       0/0 [21/28] voluntary/involuntary context switches
>         ! buffer usage stats:
>         !       Shared blocks:          0 read,          0 written,
> buffer hit rate = 0.00%
>         !       Local  blocks:          0 read,          0 written,
> buffer hit rate = 0.00%
>         !       Direct blocks:          0 read,          0 written
> STATEMENT:  SELECT
> pg_file_read('pg_log/postgresql-2010-06-30_084323.log', 0, 50000)
> LOG:  unexpected EOF on client connection
> LOG:  unexpected EOF on client connection
> LOG:  unexpected EOF on client connection
> 
> 
> 6) /var/log/messages contains this:
> Jun 30 09:43:23 localhost kernel: pgadmin3[5984]: segfault at 700000005f
> ip 000000382b815d2c sp 00007fff574b47f0 error 4 in
> libwx_gtk2u_core-2.8.so.0.7.0[382b600000+3a3000]
> Jun 30 09:43:24 localhost abrt[28553]: saved core dump of pid 5984
> (/usr/bin/pgadmin3) to /var/cache/abrt/ccpp-1277883803-5984.new/coredump
> (20774912 bytes)
> Jun 30 09:43:24 localhost abrtd: Directory 'ccpp-1277883803-5984'
> creation detected
> Jun 30 09:43:24 localhost abrtd: Crash is in database already (dup of
> /var/cache/abrt/ccpp-1277659638-8405)
> Jun 30 09:43:24 localhost abrtd: Deleting crash ccpp-1277883803-5984
> (dup of ccpp-1277659638-8405), sending dbus signal
> (coredumps available on request)
> 
> The DB runs as user postgres, which was created at DBMS install, and all
> files belong to this user. selinux is enforcing/strict.
> pgadmin3 runs as root and connects to the DBs using various roles, it is
> very stable.
> Changing the user or the permissions seems to make no difference.
> When I try to shoow the "Server Status", it pops up the window for a
> short moment and then the whole application crashes.
> 
> If I can supply any other info, please ask / send instructions.

I installed a VM with FC12, and cannot reproduce your issue. I don't
launch pgadmin3 as root, but I don't think it can have that effect.
Perhaps there is something in your log files it doesn't like. Care to
share them with me?


-- 
Guillaumehttp://www.postgresql.frhttp://dalibo.com

Hi,

It's a late reply, sry about that.
pgAdmin stalls on me when i open the server status window, since i
installed the adminpack.
(pgAdmin version 1.12.2 on ubuntu 11.04)
It shows an error, which doesn't react to click events.
from the pgadmin log:
2011-09-21 11:14:16 ERROR  : ERROR:  could not read directory
"pg_log": No such file or directory
The applications then stalls.

Postgres in on a server running postgresql 8.4 on ubuntu lucid.
The adminpack is from tha package postgresql-contrib-8.4, probably
from the lucid-updates repository, the version is 8.4.8-0ubuntu0.10.04
topmost line is /* $PostgreSQL:
pgsql/contrib/adminpack/adminpack.sql.in,v 1.6 2007/11/13 04:24:27
momjian Exp $ */

So i made a symlink /var/lib/postgresql/8.4/main/pg_log --> /var/log/postgresql/
Now pgAdmin doesn't stall anymore when i open the server status, but
the log windows shows a message:
"logging_collector not enabled or log_filename misconfigured"

I didn't change the default location of the logs.

Cheers,

WBL

On Sat, Jul 3, 2010 at 12:55 AM, Guillaume Lelarge
<guillaume@lelarge.info> wrote:
> Le 30/06/2010 10:16, r d a écrit :
>> [...]
>> pgadmin3 crashes when I try to see the server statistics (menu Tools -->
>> Server Status).
>> I think that I configured my installation correctly, details follow below.
>>
>> My configuration:
>> ---------------------------------------------------------------------------------
>> - Linux FC12 installed clean, plus YUM upgrades from network
>> uname -a == Linux L3 2.6.32.14-127.fc12.x86_64 #1 SMP Fri May 28
>> 04:30:39 UTC 2010 x86_64 x86_64 x86_64 GNU/Linux
>>
>> - PostgreSQL 8.4.4 installed via YUM from net
>> postgresql-server-8.4.4-1.fc12.x86_64
>> (and all postgresql-* packages plus dependencies)
>>
>> - pgadmin3 1.10.3 - installed via YUM from net
>> pgadmin3-1.10.3-2.fc12.x86_64
>>
>> - wx 2.8.11 installed from net via YUM
>> wxBase-2.8.11-1.fc12.x86_64
>> wxPython-2.8.11.0-1.fc12.x86_64
>> wxGTK-media-2.8.11-1.fc12.x86_64
>> wxGTK-devel-2.8.11-1.fc12.x86_64
>> wxGTK-gl-2.8.11-1.fc12.x86_64
>> wxPython-devel-2.8.11.0-1.fc12.x86_64
>> wxGTK-2.8.11-1.fc12.x86_64
>> ---------------------------------------------------------------------------------
>>
>> I set up the instrumentation, logging, ...  as per instructions, i.e.
>> 1) psql --username=username dbname -f adminpack.sql   (actually, I have
>> all contribs installed into all databases)
>> 2) logging output filename is (postgresql.conf) log_filename =
>> 'postgresql-%Y-%m-%d_%H%M%S.log'
>> 3) logs are written into data/pg_log (default)
>> 4) log files look like this (ls -l)
>>
>> -rw-------. 1 postgres postgres       171 2010-06-30 01:41
>> postgresql-2010-06-30_000000.log
>> -rw-------. 1 postgres postgres    322532 2010-06-30 09:43
>> postgresql-2010-06-30_084323.log
>>
>> 5) Sample contents of a log content looks like this:
>> STATEMENT:  SELECT
>> pg_file_read('pg_log/postgresql-2010-06-30_084323.log', 0, 50000)
>> LOG:  EXECUTOR STATISTICS
>> DETAIL:  ! system usage stats:
>>         !       0.000242 elapsed 0.000000 user 0.000000 system sec
>>         !       [0.016997 user 0.005999 sys total]
>>         !       0/0 [64/0] filesystem blocks in/out
>>         !       0/4 [0/1563] page faults/reclaims, 0 [0] swaps
>>         !       0 [0] signals rcvd, 0/0 [0/0] messages rcvd/sent
>>         !       0/0 [21/28] voluntary/involuntary context switches
>>         ! buffer usage stats:
>>         !       Shared blocks:          0 read,          0 written,
>> buffer hit rate = 0.00%
>>         !       Local  blocks:          0 read,          0 written,
>> buffer hit rate = 0.00%
>>         !       Direct blocks:          0 read,          0 written
>> STATEMENT:  SELECT
>> pg_file_read('pg_log/postgresql-2010-06-30_084323.log', 0, 50000)
>> LOG:  unexpected EOF on client connection
>> LOG:  unexpected EOF on client connection
>> LOG:  unexpected EOF on client connection
>>
>>
>> 6) /var/log/messages contains this:
>> Jun 30 09:43:23 localhost kernel: pgadmin3[5984]: segfault at 700000005f
>> ip 000000382b815d2c sp 00007fff574b47f0 error 4 in
>> libwx_gtk2u_core-2.8.so.0.7.0[382b600000+3a3000]
>> Jun 30 09:43:24 localhost abrt[28553]: saved core dump of pid 5984
>> (/usr/bin/pgadmin3) to /var/cache/abrt/ccpp-1277883803-5984.new/coredump
>> (20774912 bytes)
>> Jun 30 09:43:24 localhost abrtd: Directory 'ccpp-1277883803-5984'
>> creation detected
>> Jun 30 09:43:24 localhost abrtd: Crash is in database already (dup of
>> /var/cache/abrt/ccpp-1277659638-8405)
>> Jun 30 09:43:24 localhost abrtd: Deleting crash ccpp-1277883803-5984
>> (dup of ccpp-1277659638-8405), sending dbus signal
>> (coredumps available on request)
>>
>> The DB runs as user postgres, which was created at DBMS install, and all
>> files belong to this user. selinux is enforcing/strict.
>> pgadmin3 runs as root and connects to the DBs using various roles, it is
>> very stable.
>> Changing the user or the permissions seems to make no difference.
>> When I try to shoow the "Server Status", it pops up the window for a
>> short moment and then the whole application crashes.
>>
>> If I can supply any other info, please ask / send instructions.
>
> I installed a VM with FC12, and cannot reproduce your issue. I don't
> launch pgadmin3 as root, but I don't think it can have that effect.
> Perhaps there is something in your log files it doesn't like. Care to
> share them with me?
>
>
> --
> Guillaume
>  http://www.postgresql.fr
>  http://dalibo.com
>
> --
> Sent via pgadmin-support mailing list (pgadmin-support@postgresql.org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgadmin-support
>



--
"Patriotism is the conviction that your country is superior to all
others because you were born in it." -- George Bernard Shaw

On Wed, 2011-09-21 at 12:24 +0200, Willy-Bas Loos wrote:
> [...]
> It's a late reply, sry about that.
> pgAdmin stalls on me when i open the server status window, since i
> installed the adminpack.
> (pgAdmin version 1.12.2 on ubuntu 11.04)
> It shows an error, which doesn't react to click events.
> from the pgadmin log:
> 2011-09-21 11:14:16 ERROR  : ERROR:  could not read directory
> "pg_log": No such file or directory
> The applications then stalls.
> 
> Postgres in on a server running postgresql 8.4 on ubuntu lucid.
> The adminpack is from tha package postgresql-contrib-8.4, probably
> from the lucid-updates repository, the version is 8.4.8-0ubuntu0.10.04
> topmost line is /* $PostgreSQL:
> pgsql/contrib/adminpack/adminpack.sql.in,v 1.6 2007/11/13 04:24:27
> momjian Exp $ */
> 
> So i made a symlink /var/lib/postgresql/8.4/main/pg_log --> /var/log/postgresql/
> Now pgAdmin doesn't stall anymore when i open the server status, but
> the log windows shows a message:
> "logging_collector not enabled or log_filename misconfigured"
> 
> I didn't change the default location of the logs.
> 

The adminpack should only work if you use the logging_collector and the
default values for log_directory and log_filename. One of these is
probably wrongly configured.


-- 
Guillaume http://blog.guillaume.lelarge.info http://www.dalibo.com

On Fri, Sep 23, 2011 at 10:07 PM, Guillaume Lelarge
<guillaume@lelarge.info> wrote:
> The adminpack should only work if you use the logging_collector and the
> default values for log_directory and log_filename. One of these is
> probably wrongly configured.

the logging collector is off by default, yet the log is being collected.
this is on ubuntu server lucid (10.04, the current LTS version).

Maybe it would be a good idea to run show logging_collector
upon installing the adminpack, and raise a warning notice if it is 'off'.

too bad you can only use this feature with the logging_collector. I
won't change the logging now that i am content with it (apart from
reading it in pgAdmin), I think that many ubuntu users will feel the
same.
Also, pgAdmin crashes (no, stalls) on the deafult ubuntu server
configuration with the adminpack installed, unless you make the
symlink (which was an educated guess). I think that that should at
least be fixed.

Cheers,

WBL



-- 
"Patriotism is the conviction that your country is superior to all
others because you were born in it." -- George Bernard Shaw

On Mon, 2011-09-26 at 12:02 +0200, Willy-Bas Loos wrote:
> On Fri, Sep 23, 2011 at 10:07 PM, Guillaume Lelarge
> <guillaume@lelarge.info> wrote:
> > The adminpack should only work if you use the logging_collector and the
> > default values for log_directory and log_filename. One of these is
> > probably wrongly configured.
> 
> the logging collector is off by default, yet the log is being collected.

I may be wrong there. I'll have to check when I find some time to work
on it... sigh...

> this is on ubuntu server lucid (10.04, the current LTS version).
> 
> Maybe it would be a good idea to run
>   show logging_collector
> upon installing the adminpack, and raise a warning notice if it is 'off'.
> 
> too bad you can only use this feature with the logging_collector. I
> won't change the logging now that i am content with it (apart from
> reading it in pgAdmin), I think that many ubuntu users will feel the
> same.

Well, I can't complain. There are many things I would like to see
happening on the log files, but it sure won't be easy to fix.

> Also, pgAdmin crashes (no, stalls) on the deafult ubuntu server
> configuration with the adminpack installed, unless you make the
> symlink (which was an educated guess). I think that that should at
> least be fixed.

I guess so. But we'll have to wait till I find time to install ubuntu on
a VM, and debug from it. Unless someone beats me to it.


-- 
Guillaume http://blog.guillaume.lelarge.info http://www.dalibo.com

On Mon, 2011-09-26 at 12:02 +0200, Willy-Bas Loos wrote:
> On Fri, Sep 23, 2011 at 10:07 PM, Guillaume Lelarge
> <guillaume@lelarge.info> wrote:
> Also, pgAdmin crashes (no, stalls) on the deafult ubuntu server
> configuration with the adminpack installed, unless you make the
> symlink (which was an educated guess). I think that that should at
> least be fixed.
> 

It didn't happen to me.

Here is what I did:
* download ubuntu 11.04, 64 bits
* install it
* launch it, and connect it as single user
* sudo apt-get install aptitude
* sudo aptitude install postgresql-8.4 postgresql-client-8.4
* sudo aptitude install postgresql-contrib-8.4
* sudo aptitude install postgresql-server-dev-8.4 libpq-dev
* sudo aptitude install libwxgtk2.8-dbg libwxgtk2.8-dev
* sudo aptitude install libxml2-dev libxslt1-dev
* download pgadmin3 1.14 source package
* tar xvfz Downloads.pgadmin3-1.14.0.tar.gz
* cd pgadmin3-1.14.0
* ./configure --enable-debug
* sudo make
* sudo make install
* vim /etc/postgresql/8.4/main/postgresql.conf listen_addresses = '*'
* vim /etc/postgresql/8.4/main/pg_hba.conf (change the authentification method to trust... yeah, I know, that's bad :)
)
* sudo /etc/init.d/postgresql restart
* psql -U postgres -f /usr/share/postgresql/8.4/contrib/adminpack.sql
postgres
* /usr/local/pgadmin3/bin/pgadmin3 * add new server * connection to new server * launch server status   ... and no
crash  ... just a single message telling me "could not read directory       pg_log"
 


-- 
Guillaume http://blog.guillaume.lelarge.info http://www.dalibo.com