'lti' USER sessions filling all available connections into RepServer

[Former Member]

Sometime, Replication Agent connections are  opening  USER Sessions into Replication Server

Each retry is adding a new USER Session. gradually filling up all the limit of 'num_client_connections' Parameter

[Former Member]

Hi Denis,

is this a continuation of the same issue you had at the beginning of the month?

It looked like you knew how to reproduce this issue then?

I would recommend to repeat Mark's suggestions:

At this point I'd suggest:

- you get 1 of the questionable repagents reconfigured and then stop it (ie, leave configured but disconnected from repserver) [please also provide details on how you reconfigured the repagent, eg, did you manually configure the repagent or use rs_init?]

- bounce the repserver, start it with a new errorlog file

- start the repagent

- run 'sp_help_rep_agent'

- if the repagent is not connecting to the repserver ... wait about 5 minutes and then continue with the following ...

- provide the complete repserver errorlog file

- provide all lines from the ASE errrolog from just prior to starting the repagent

I would prefer to have the whole ASE errorlog from the latest start up to the current end.)

- provide all lines from your repserver auditing file from just prior to starting the repagent

- provide a snapshot of 'admin who'

The objective is to try and troubleshoot one of your problematic repagents (ie, one of the repagents that cannot connect to the repserver).

Many thanks,

- Jon

[Former Member]

Hi Jon,

as you can see next,

it is now easy to me reproducing that situation.

Where a 'Primary Server.Database' as 'RepAgent' running and 'log transfer'at 'on'

I just have to:

1) suspend log transfer

2) stop RepAgent

3) re-start RepAgent

at that time, 'error '14039' from RePAgent will be displayed one time onto 'Primary Server.Database' errorlog

and if 'RepServer Audit' is at 'yes' it will show a 'failure' message one time per minute (default RepAgent connection cycle).

For each new failure, a 'Spid RepServer Session will be created and never resolved.

If no action is take, then by the time, a Saturation of the maximum available 'Sessions' will happen, making not possible connection to RepServer !!!

When 'resuming log transfer', failure errors are no more displayed into Audit and everything is going to normal.

Unfortunately,

the accumulated and 'pending Sessions' in RepServer never disappear !!!

We did not find mention of that in 'Ebf' released, so we will open a Case at Sybase.

Kinds regards,
Denis

admin who

97 REP AGENT  Awaiting Command   
SYBDEV07.DBO_DEV_DB77

suspend log transfer from SYBDEV07.DBO_DEV_DB77

Suspending
LogTransfer for SYBDEV07.DBO_DEV_DB77

1. I.
   2013/09/16 08:20:53. AUDIT: incoming command (issued by as_rep): suspend log
   transfer from SYBDEV07.DBO_DEV_DB77

admin who

97 REP AGENT
Suspended          
SYBDEV07.DBO_DEV_DB77

waiting
5 minutes . . .

no
error message

no
'unresolved lti connetion' into RepServer

rrR -SSYBDEV07 -q"exec
DBO_DEV_DB77..sp_stop_rep_agent DBO_DEV_DB77"

The
Replication Agent thread for database 'DBO_DEV_DB77' is being stopped.

00:0007:00000:00043:2013/09/16
08:33:10.90 server  RepAgent(13): Received the following error message
from the Replication Server: Msg 14040. Replication Agent for
SYBDEV07.DBO_DEV_DB77 has been Suspended. Disconnecting Replication Agent..

00:0008:00000:00043:2013/09/16
08:33:11.63 server  Shutting down Rep Agent for database, 'DBO_DEV_DB77'
(dbid = 13).

1. E.
   2013/09/16 08:31:14. ERROR #14040 REP AGENT(SYBDEV07.DBO_DEV_DB77) -
   /execint.c(1485)

      
Replication Agent for SYBDEV07.DBO_DEV_DB77 has been Suspended. Disconnecting
Replication Agent.

rrR -SSYBDEV07 -q"exec
DBO_DEV_DB77..sp_start_rep_agent DBO_DEV_DB77"

Replication
Agent thread is started for database 'DBO_DEV_DB77'.

00:0007:00000:00319:2013/09/16
08:35:54.25 server  Started Rep Agent on database, 'DBO_DEV_DB77' (dbid =
13).

00:0007:00000:00319:2013/09/16
08:35:54.45 server  RepAgent(13): Received the following error message from the Replication
Server: Msg 14039. LogTransfer for SYBDEV07.DBO_DEV_DB77 is Suspended.

admin who

187
USER       Awaiting
Command     SRSSTA01_ra

188
USER       Awaiting
Command     SRSSTA01_ra

189
USER       Awaiting
Command     SRSSTA01_ra

1. I.
   2013/09/16 08:33:57. AUDIT The following command batch has one or more
   failures:
2. I.
   2013/09/16 08:33:58. AUDIT: incoming command (issued by SRSSTA01_ra): connect
   source lti "SYBDEV07"."DBO_DEV_DB77" 750
   passthru
3. I.
   2013/09/16 08:34:57. AUDIT The following command batch has one or more
   failures:
4. I.
   2013/09/16 08:34:57. AUDIT: incoming command (issued by SRSSTA01_ra): connect source lti
   "SYBDEV07"."DBO_DEV_DB77" 750 passthru
5. I.
   2013/09/16 08:35:57. AUDIT The following command batch has one or more
   failures:
6. I.
   2013/09/16 08:35:57. AUDIT: incoming command (issued by SRSSTA01_ra): connect
   source lti "SYBDEV07"."DBO_DEV_DB77" 750
   passthru
7. I.
   2013/09/16 08:36:33. AUDIT: incoming command (issued by as_rep): resume log
   transfer from SYBDEV07.DBO_DEV_DB77

resume log transfer from SYBDEV07.DBO_DEV_DB77

Resuming
LogTransfer for SYBDEV07.DBO_DEV_DB77

1. I.
   2013/09/16 08:36:33. AUDIT: incoming command (issued by as_rep): resume log
   transfer from SYBDEV07.DBO_DEV_DB77

1. I.
   2013/09/16 08:36:57. Replication Agent for SYBDEV07.DBO_DEV_DB77 connected in
   passthru mode.

14039

Log   Transfer for LTM for RS_source_ds. RS_source_db is suspended.   The connect source is refused because Log Transfer is suspended.

LTM   for MVS shuts down regardless of Stop_on_error value.