On Tue, Jan 18, 2011 at 11:55 AM, Daniel-Constantin Mierla <miconda@gmail.com> wrote:

Hello,

I did a new tuning, skipping reconnect in submit query if connection is ok - the result might be broken due to other reasons and will be reported when fetching it. If you can try again and report, would be appreciated.

Thanks,
Daniel
On 1/18/11 11:34 AM, Javier Gallart wrote:
Hi Daniel

I tested the patch but the mtree module fails to start:
Jan 18 04:15:55 kam-test ./kamailio[4616]: DEBUG: <core> [db.c:202]: using db bind api for db_postgres
Jan 18 04:15:55 kam-test ./kamailio[4616]: DEBUG: <core> [db.c:285]: connection 0x931768 not found in pool
Jan 18 04:15:55 kam-test ./kamailio[4616]: DEBUG: db_postgres [km_pg_con.c:51]: db_id = 0x931768
Jan 18 04:15:55 kam-test ./kamailio[4616]: DEBUG: db_postgres [km_pg_con.c:64]: 0x9319b8=pkg_malloc(72)
Jan 18 04:15:55 kam-test ./kamailio[4616]: DEBUG: db_postgres [km_pg_con.c:76]: opening connection: postgres://xxxx:xxxx@79.170.68.171/routing
Jan 18 04:15:55 kam-test ./kamailio[4616]: DEBUG: db_postgres [km_pg_con.c:80]: PQsetdbLogin(0x2259780)
Jan 18 04:15:55 kam-test ./kamailio[4616]: DEBUG: mtree [mtree_mod.c:255]: database connection opened successfully
Jan 18 04:15:55 kam-test ./kamailio[4616]: WARNING: db_postgres [km_dbase.c:161]: postgres result check failed with code 7 (PGRES_FATAL_ERROR)
Jan 18 04:15:55 kam-test ./kamailio[4616]: WARNING: db_postgres [km_dbase.c:165]: postgres query command failed, connection status 0, error []
Jan 18 04:15:55 kam-test ./kamailio[4616]: WARNING: db_postgres [km_dbase.c:165]: postgres query command failed, connection status 0, error [another command is already in progress#012]
Jan 18 04:15:55 kam-test ./kamailio[4616]: WARNING: db_postgres [km_dbase.c:165]: postgres query command failed, connection status 0, error [another command is already in progress#012]
Jan 18 04:15:55 kam-test ./kamailio[4616]: ERROR: db_postgres [km_dbase.c:173]: 0x931660 PQsendQuery Error: another command is already in progress#012 Query: select tname,tprefix,tvalue from mtree order by tname
Jan 18 04:15:55 kam-test ./kamailio[4616]: ERROR: <core> [db_query.c:101]: error while submitting query
Jan 18 04:15:55 kam-test ./kamailio[4616]: ERROR: mtree [mtree_mod.c:626]: Error while querying db
Jan 18 04:15:55 kam-test ./kamailio[4616]: ERROR: mtree [mtree_mod.c:296]: cannot load trees from database
Jan 18 04:15:55 kam-test ./kamailio[4616]: DEBUG: <core> [db_pool.c:102]: removing connection from the pool
Jan 18 04:15:55 kam-test ./kamailio[4616]: DEBUG: db_postgres [km_pg_con.c:124]: PQfinish(0x2259780)
Jan 18 04:15:55 kam-test ./kamailio[4616]: DEBUG: db_postgres [km_pg_con.c:128]: pkg_free(0x9319b8)
Jan 18 04:15:55 kam-test ./kamailio[4616]: ERROR: <core> [sr_module.c:875]: init_mod(): Error while initializing module mtree (/usr/local/kamailio.20110117/lib/kamailio/modules/mtree.so)

I don't see an easy way to fix this, because even after the server restart, PQsendQuery will return 1, and therefore db_postgres_fetch_result is executed and fails. I will keep looking into this, for now it's not a big issue since we know that the second attempt works fine....

Thanks

Javi
On Sun, Jan 16, 2011 at 11:09 PM, Daniel-Constantin Mierla <miconda@gmail.com> wrote:
Hello,

made a commit to fix in such case, when the broken connection is reported after doing actually the query, attempting to reconnect then. Let me know if works now.

Cheers,
Daniel
On 1/16/11 6:17 PM, Javier Gallart wrote:
Hi Daniel,

I'm using postgres 8.4 at the server and libpq5 library for the client. I've been testing again and the problem is that after a db restart (no matter if done through a service restart or having t down for some time), the PQstatus(CON_CONNECTION(_con)) function returns a CONNECTION_OK and therefore it doesn't try to reconnect. It's on the second attempt of the mt_reload that it returns the correct CONNECTION_BAD value, and reconnection at that point works perfectly well.

First attempt after db restart (I've added a dbg message if we get a CONNECTION_OK):

Jan 16 12:03:00 kam-test ./kamailio[22214]: DEBUG: db_postgres [km_dbase.c:129]: connection OK
Jan 16 12:03:00 kam-test ./kamailio[22214]: DEBUG: db_postgres [km_dbase.c:155]: sending query ok: 0x92f888 - [select tname,tprefix,tvalue from mtree order by tname]
Jan 16 12:03:00 kam-test ./kamailio[22214]: DEBUG: <core> [db_res.c:118]: allocate 48 bytes for result set at 0x94cd08
Jan 16 12:03:00 kam-test ./kamailio[22214]: DEBUG: db_postgres [km_dbase.c:222]: 0x92f888 PQresultStatus(PGRES_FATAL_ERROR) PQgetResult(0x1a30d10)
Jan 16 12:03:00 kam-test ./kamailio[22214]: ERROR: db_postgres [km_dbase.c:238]: 0x92f888 - invalid query, execution aborted
Jan 16 12:03:00 kam-test ./kamailio[22214]: ERROR: db_postgres [km_dbase.c:239]: 0x92f888 - PQresultStatus(PGRES_FATAL_ERROR)
Jan 16 12:03:00 kam-test ./kamailio[22214]: ERROR: db_postgres [km_dbase.c:240]: 0x92f888: server closed the connection unexpectedly#012#011This probably means the server terminated abnormally#012#011before or while processing the request.#012

Second attempt:

Jan 16 12:03:08 kam-test ./kamailio[22214]: DEBUG: db_postgres [km_dbase.c:132]: connection reset
Jan 16 12:03:08 kam-test ./kamailio[22214]: DEBUG: db_postgres [km_dbase.c:315]: PQclear(0x1a30d10) result set
Jan 16 12:03:08 kam-test ./kamailio[22214]: DEBUG: db_postgres [km_dbase.c:155]: sending query ok: 0x92f888 - [select tname,tprefix,tvalue from mtree order by tname]

Thanks!

Javi
On Fri, Jan 14, 2011 at 11:42 AM, Daniel-Constantin Mierla <miconda@gmail.com> wrote:
Hello,

what version of postgres and postgres client library are you using?

According to the docs from devel manual, this is the function that should be used to reconnect:

PQreset
Resets the communication channel to the server.
void PQreset(PGconn *conn);
This function will close the connection to the server and attempt to reestablish a new connection to the same server, using all the same parameters previously used. This may be useful for error recovery if a working connection is lost.
You may see more here:
http://www.network-theory.co.uk/docs/postgresql/vol2/DatabaseConnectionControlFunctions.html

Somehow seems not to work in this case. Can you try also with a shorter down time for postgres server, like a normal restart? Would be intrested to see if the client side loses somehow connection parameters due to long time unconnected.

Cheers,
Daniel
On 1/14/11 11:15 AM, Javier Gallart wrote:
Hi Daniel

this is my test:

1- Kamailio starts and correctly loads the mtree
2- I make whatever modification in the tree and reload with mt_reload
Jan 14 04:19:06 kam-test ./kamailio[31205]: DEBUG: db_postgres [km_dbase.c:153]: sending query ok: 0x92f880 - [select tname,tprefix,tvalue from mtree order by tname]
Jan 14 04:19:06 kam-test ./kamailio[31205]: DEBUG: <core> [db_res.c:118]: allocate 48 bytes for result set at 0x94cd00
Jan 14 04:19:06 kam-test ./kamailio[31205]: DEBUG: db_postgres [km_dbase.c:220]: 0x92f880 PQresultStatus(PGRES_TUPLES_OK) PQgetResult(0x29f05f0)
Jan 14 04:19:06 kam-test ./kamailio[31205]: DEBUG: db_postgres [km_res.c:108]: 3 columns returned from the query
Jan 14 04:19:06 kam-test ./kamailio[31205]: DEBUG: <core> [db_res.c:155]: allocate 24 bytes for result names at 0x94cc60
[....]
Jan 14 04:19:06 kam-test ./kamailio[31205]: DEBUG: db_postgres [km_res.c:267]: freeing row buffer at 0x94c940
Jan 14 04:19:06 kam-test ./kamailio[31205]: DEBUG: mtree [mtree.c:738]: adding new tname [1]
Jan 14 04:19:06 kam-test ./kamailio[31205]: DEBUG: <core> [db_res.c:54]: freeing 3 rows
[...]
Jan 14 04:19:06 kam-test ./kamailio[31205]: DEBUG: <core> [db_res.c:94]: freeing result names at 0x94cc60
Jan 14 04:19:06 kam-test ./kamailio[31205]: DEBUG: <core> [db_res.c:99]: freeing result types at 0x94cbc0
Jan 14 04:19:06 kam-test ./kamailio[31205]: DEBUG: <core> [db_res.c:136]: freeing result set at 0x94cd00
Jan 14 04:19:06 kam-test ./kamailio[31205]: DEBUG: db_postgres [km_dbase.c:313]: PQclear(0x29f05f0) result set

3. Postgres DB restarts (either through restart, or /stop/wait a minute/start)
No log is written in Kamailio when this happens.

4.- I try again to reload the tree, and these are the logs:
Jan 14 04:21:42 kam-test ./kamailio[31205]: DEBUG: db_postgres [km_dbase.c:153]: sending query ok: 0x92f880 - [select tname,tprefix,tvalue from mtree order by tname]
Jan 14 04:21:42 kam-test ./kamailio[31205]: DEBUG: <core> [db_res.c:118]: allocate 48 bytes for result set at 0x94cd00
Jan 14 04:21:42 kam-test ./kamailio[31205]: DEBUG: db_postgres [km_dbase.c:220]: 0x92f880 PQresultStatus(PGRES_FATAL_ERROR) PQgetResult(0x29ef0a0)
Jan 14 04:21:42 kam-test ./kamailio[31205]: ERROR: db_postgres [km_dbase.c:236]: 0x92f880 - invalid query, execution aborted
Jan 14 04:21:42 kam-test ./kamailio[31205]: ERROR: db_postgres [km_dbase.c:237]: 0x92f880 - PQresultStatus(PGRES_FATAL_ERROR)
Jan 14 04:21:42 kam-test ./kamailio[31205]: ERROR: db_postgres [km_dbase.c:238]: 0x92f880: server closed the connection unexpectedly#012#011This probably means the server terminated abnormally#012#011before or while processing the request.#012
Jan 14 04:21:42 kam-test ./kamailio[31205]: DEBUG: <core> [db_res.c:81]: freeing 0 columns
Jan 14 04:21:42 kam-test ./kamailio[31205]: DEBUG: <core> [db_res.c:136]: freeing result set at 0x94cd00
Jan 14 04:21:42 kam-test ./kamailio[31205]: ERROR: mtree [mtree_mod.c:633]: Error while fetching result
Jan 14 04:21:42 kam-test ./kamailio[31205]: ERROR: db_postgres [km_dbase.c:329]: invalid parameter value
Jan 14 04:21:42 kam-test ./kamailio[31205]: ERROR: mtree [mtree_mod.c:744]: cannot re-load info from database

Thanks in advance

Regards

Javi
On Thu, Jan 13, 2011 at 1:56 PM, Daniel-Constantin Mierla <miconda@gmail.com> wrote:
Can you send the log messages (debug=3) for that moment (restart of pg server and making a new call/registration)?

Daniel

On 1/13/11 12:00 PM, Javier Gallart wrote:
Hi Daniel

I have just tested it and I'm getting the same results. This is the kamailio version I'm testing:
root@kam-test:/usr/local/kamailio/sbin# ./kamailio -V
version: kamailio 3.2.0-dev2 (x86_64/linux) 35568f

Thanks!

Javi

On Tue, Jan 4, 2011 at 3:46 PM, Javier Gallart <jgallartm@gmail.com> wrote:

Hi list

we have been testing the mtree module. We do periodical reloads of the mtree table with the "mi mt_reload" command through xmlrpc. It works fine except when the db is restarted. If we manually restart the db and then try the mt_reload commands we get this error:

Jan 4 09:13:17 server /usr/local/kamailio-3.1.1/sbin/kamailio[28534]: ERROR: db_postgres [km_dbase.c:152]: 0xa6a490 PQsendQuery Error: server closed the connection unexpectedly#012#011This probably means the server terminated abnormally#012#011before or while processing the request.#012 Query: select partition,prefix,route from trees order by partition

Jan 4 09:13:17 server /usr/local/kamailio-3.1.1/sbin/kamailio[28534]: ERROR: <core> [db_query.c:101]: error while submitting query

Jan 4 09:13:17 server /usr/local/kamailio-3.1.1/sbin/kamailio[28534]: ERROR: mtree [mtree_mod.c:622]: Error while querying db

Jan 4 09:13:17 server /usr/local/kamailio-3.1.1/sbin/kamailio[28534]: ERROR: mtree [mtree_mod.c:738]: cannot re-load info from database

Is there a way to reconnect to the db before sending that command?

Thanks in advance

Regards

Javi
_______________________________________________
SIP Express Router (SER) and Kamailio (OpenSER) - sr-users mailing list
sr-users@lists.sip-router.org
http://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-users
-- 
Daniel-Constantin Mierla
Kamailio (OpenSER) Advanced Training
Jan 24-26, 2011, Irvine, CA, USA
http://www.asipto.com
_______________________________________________
SIP Express Router (SER) and Kamailio (OpenSER) - sr-users mailing list
sr-users@lists.sip-router.org
http://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-users
-- 
Daniel-Constantin Mierla
Kamailio (OpenSER) Advanced Training
Jan 24-26, 2011, Irvine, CA, USA
http://www.asipto.com
-- 
Daniel-Constantin Mierla
Kamailio (OpenSER) Advanced Training
Jan 24-26, 2011, Irvine, CA, USA
http://www.asipto.com
_______________________________________________
SIP Express Router (SER) and Kamailio (OpenSER) - sr-users mailing list
sr-users@lists.sip-router.org
http://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-users
-- 
Daniel-Constantin Mierla
Kamailio (OpenSER) Advanced Training
Jan 24-26, 2011, Irvine, CA, USA
http://www.asipto.com