ORA 3136 Errors in Oracle Alert log and chronic performance

  • 25 November 2021
  • 1 reply

Userlevel 1
Badge +4

We keep getting issues with IFS where the application grinds to almost a stop. Tasks that should take a few seconds take 10minutes and users can no longer log on. When we look in the Oracle alert log we see lots of occurrences of ORA 3136. Episodes like this seem to last about one hour then the application returns to normal use. We noticed that memory utilization on one or other of the extended servers  shoots up to 100% during the onset of these events but we cannot find any reason for this. It tends to happen at about 9am and 4pm when there is more login/logout activity and a bit around lunch time ~noon but can see no real cause. It doesn’t happen everyday. It was very bad last Friday and Monday of this week. Since then it’s performed well . Has anyone seen anything similar please? 

1 reply

Userlevel 7
Badge +13

Hi @pnaughton,

According to Oracle Metalink documents, the reason for the ORA-3136 error being thrown in the alert log is when a client fails to complete the authentication process in the time allowed by INBOUND_CONNECT_TIMEOUT parameter. 

  1. Very often this error is thrown because a client or application is attempting to connect using the wrong username or password.
  2. The server receives a valid client connection request but the client takes a long time to authenticate more than the default 60 seconds.
  3. The DB server is heavily loaded due to which it cannot finish the client logon within the timeout specified.

You could check the Middleware Server logs to see if you can find any invalid username/password errors corresponding to these Oracle errors and try to identify which accounts could be causing it. Also look out for errors that say “Abandoning transaction after xx,xxx seconds”.

You could also look into additional installations such as Crystal Web Service or IFS Print Agent and see if there are any installations with incorrect passwords for IFS System accounts such as IFSPRINT/IFSSYS.

Furthermore, it would be a good idea to take thread dumps from each of the Middleware Servers to see if you can find something in them as well, since you have mentioned you have seen CPU spikes on extended servers as well.

If you fail to find out much from above, if you have Oracle ASFU support through IFS, I would recommend you to open a case and report this issue so that we can open an Oracle SR and request their assistance. 

If you have access to My Oracle Support, then you can refer your DBA to check out following Oracle documents regarding this:

  • Troubleshooting Guide ORA-3136: WARNING Inbound Connection Timed Out (Doc ID 465043.1)
  • ORA-1017 May Cause ORA-3136 WARNING : Inbound Connection Timed Out in Alert Log (Doc ID 793259.1)

If you open a case, we can provide the information included in these documents as well. 

Hope this helps!