cancel
Showing results for 
Search instead for 
Did you mean: 

DBstudio no connection: Connection reset by peer

Former Member
0 Kudos

While trying to connect to a DB with DBStudio (7.9.08.30) I keep getting the error: Connection T3812 could not send data [Connection reset by peer: socket write error].

Detail:

com.sap.sdb.tools.usermanagement.UM_Exception: Connection T3812 could not send data [Connection reset by peer: socket write error].

    at com.sap.sdb.tools.usermanagement.UM_SessionDBM.execute(UM_SessionDBM.java:70)

    at com.sap.sdb.tools.databasemanagement.internal.DM_Session.execute(DM_Session.java:282)

    at com.sap.sdb.tools.databasemanagement.internal.commands.DM_BaseCommand.doExecute(DM_BaseCommand.java:187)

    at com.sap.sdb.tools.databasemanagement.internal.commands.DM_DatabaseCommand.doExecute(DM_DatabaseCommand.java:55)

    at com.sap.sdb.tools.databasemanagement.internal.commands.DM_BaseCommand.execute(DM_BaseCommand.java:119)

    at com.sap.sdb.tools.databasemanagement.internal.DM_DatabaseTrace.refresh(DM_DatabaseTrace.java:181)

    at com.sap.sdb.tools.ui.dbm.editors.pages.RefreshablePage$1.run(RefreshablePage.java:173)

    at com.sap.sdb.tools.dbm.jobs.DBMJob$1.run(DBMJob.java:81)

At the same time the following line got written in DBM.prt:

2 DBMSrvCo A DBM Server client connection was released (client process has process ID 1100 on computer

The DBStudio is on the same Server as the DB. The firewall is switched off.

The DB parameter session_timeout is set to 0.

It looks like the DB Kernel actively stops interrupts communitcation.....

Any ideas?

Thank you,

     julian

Accepted Solutions (0)

Answers (1)

Answers (1)

JamesZ
Advisor
Advisor
0 Kudos

Hi Julian,

Can you give me the screenshots about this problem?

What is the OS?

Are you able to connect database via below command?

dbmcli -d <dbsid> -u <username>,<password> db_state

Best regards,
James

Former Member
0 Kudos

Hi James,

Yes, I can connect via dbmcli.

dbmcli on HED>db_state

OK

State

ADMIN

The OS is Windows Server 2008 R2

Best regards,

     julian

JamesZ
Advisor
Advisor
0 Kudos

Hi Julian,

It seems the problem is not a simply connect problem.

Does this issue only occur in backup list page?

Best regards,
James

Former Member
0 Kudos

Hello,

I think I found out something intreresting about this problem:

Currently I'm doing a log recovery from dbmcli:

dbmcli on HED>recover_start LOG log 11237 until 20150423 090000

While this is running, I got no chance to connect to the db. Neither with DBStudio, nor with another dbmcli session. The problem seems to be that only ONE connection at a time is possible.

Is that normal?

Thank you,

     julian

JamesZ
Advisor
Advisor
0 Kudos

Hi Julian,

I think it should be normal. Since the db is doing recover, what else we can do with the database? So it should be ok to forbidden the further log on.

Please wait for the recovery to finish and then check if there is any problem for database studio to connect database.

Best regards,
James

Former Member
0 Kudos

Hello James,

I think there's a bigger problem because even the restore command is not coming back. The resore should be finished already.

But there's no chance to see if the restore ist doing anything.

Regards,

  julian

JamesZ
Advisor
Advisor
0 Kudos

Hi Julian,

You can use os command "x_cons <dbsid> show active" to understand what the database is doing and if the restore process is doing.

Best regards,
James

Former Member
0 Kudos

Hi James,

Here's the output:

SERVERDB: HED

ID   UKT  Win   TASK       APPL Current        Timeout/ Region     Wait

          tid   type        pid state          Priority cnt try    item

T105   8  0xE60 User        872 IO Wait (R)             10    15        23(r)

I guess the system is doing nothing.

I've opened a SAP call for the problem now.

Best regards,

  julian

JamesZ
Advisor
Advisor
0 Kudos

Hi Julian,

Yes, it looks like database is doing nothing. Please shutdown the database and bring it to admin mode and do the recovery again.

This time you can use this method to monitor the restore process.

Best regards,
James

Former Member
0 Kudos

Hello,

I restarted the DB into admin mode and repeated the recovery, but still the same situation..

C:\Users\hedadm>x_cons HED show active

SERVERDB: HED

ID   UKT  Win   TASK       APPL Current        Timeout/ Region     Wait

          tid   type        pid state          Priority cnt try    item

T105   8  0x600 User       1824 IO Wait (R)             10    15        398(r)

Best regards,

  julian

JamesZ
Advisor
Advisor
0 Kudos

Hi Julian,


Please attach diagpack.tgz from maxdb run directory after executing os command below:

dbmcli -d <dbsid> -u <username>,<password> diag_pack


Best regards,
James

Former Member
0 Kudos

Hi James,

Here's the file ...

I've changed the extension from tgz to txt in order to be able to upload the file.

Best regards,

  julian

JamesZ
Advisor
Advisor
0 Kudos

Hi Julian,

I can see the log information "RESTORE log CANCEL", can you let me know how you performed restore/recovery? with database studio?


What exact error occurs before restore log cancel?

Best regards,
James