2

Oracle kills sessions remotely open at the specific time. It means if I open a session at local machine, it is ok. In my case, at every 58 minute.

For example, at 13:58, 14:58, 15:58 and so on.

It means even if I open a session at 13:57, it will be closed 1 minute after.

My environment is:

  • I tested with sqlplus and my python application.
  • DB is placed in Windows Server 2012.
  • Firewall is turned off.
  • Network is connected over local router.

I also checked the Oracle alert.log, but there is nothing special.

Any suggestions?

--------------------Edit--------------------

I get a trace logs after setting TRACE_LEVEL_SERVER on sqlnet.ora.

There seems an error but I don't know what it means.

I run two program, one is my app and the other is seqlplus.

They missed connection at the same time 17:36.

Below are the trace log from each.

[07-4월 -2016 17:36:48:829] ntt2err: soc 4160 error - operation=5, ntresnt[0]=530, ntresnt[1]=53, ntresnt[2]=0
[07-4월 -2016 17:36:48:829] nserror: nsres: id=0, op=68, ns=12570, ns2=12560; nt[0]=530, nt[1]=53, nt[2]=0; ora[0]=0, ora[1]=0, ora[2]=0
[07-4월 -2016 17:36:48:829] nioqrc: recv: packet reader error -> translated to IFCR_EOF
[07-4월 -2016 17:36:48:829] nstimarmed: no timer allocated
[07-4월 -2016 17:36:48:829] nsclose: closing transport
[07-4월 -2016 17:36:48:829] nsclose: global context check-out (from slot 0) 
complete

.

[07-4월 -2016 17:36:48:829] ntt2err: soc 2216 error - operation=5, ntresnt[0]=530, ntresnt[1]=53, ntresnt[2]=0
[07-4월 -2016 17:36:48:829] nserror: nsres: id=0, op=68, ns=12570, ns2=12560; nt[0]=530, nt[1]=53, nt[2]=0; ora[0]=0, ora[1]=0, ora[2]=0
[07-4월 -2016 17:36:48:829] nioqrc: recv: packet reader error -> translated to IFCR_EOF
[07-4월 -2016 17:36:48:829] nstimarmed: no timer allocated
[07-4월 -2016 17:36:48:829] nsclose: closing transport
[07-4월 -2016 17:36:48:829] nsclose: global context check-out (from slot 0) complete
5
  • 2
    Definitely an external influence at play here....
    – Philᵀᴹ
    Commented Mar 11, 2016 at 7:06
  • When did this start happening? Have you looked at the "at" command?
    – Vérace
    Commented Mar 11, 2016 at 11:36
  • Are there any errors in the alert log or listener logs during the time when the sessions are getting disconnected?
    – andba77
    Commented Mar 11, 2016 at 16:28
  • 1
    Maybe more of a post for the superuser forum?
    – Vérace
    Commented Mar 13, 2016 at 15:27
  • It is incredibly unlikely that Oracle is killing the session. Something else is causing the session to be dropped. I would wager that something at the networking level is causing the issue. Commented Mar 14, 2016 at 6:19

1 Answer 1

-1

If the sessions can no longer be reached, this sounds a lot like Oracle Dead Connection Detection.

From that paper:

Extract

1
  • That does not match the description of the error. The OP says that this happens at specific times of day.
    – miracle173
    Commented Mar 14, 2016 at 6:31

Not the answer you're looking for? Browse other questions tagged or ask your own question.