Forum Stats

  • 3,815,385 Users
  • 2,259,012 Discussions
  • 7,893,063 Comments

Discussions

Getting "unchecked exception" warning after upgrading to ORDS 21.4

<Feb 3, 2022 2:56:20,015 PM CST> <Warning> <oracle.ucp> <BEA-000000> <:::got unchecked exception from user-defined connection labeling callback>

<Feb 3, 2022 2:56:20,391 PM CST> <Warning> <oracle.ucp> <BEA-000000> <:::got unchecked exception from user-defined connection labeling callback>

<Feb 3, 2022 2:56:20,943 PM CST> <Warning> <oracle.ucp> <BEA-000000> <:::got unchecked exception from user-defined connection labeling callback>

In diagnostic Log

[2022-02-03T14:56:20.015-06:00] [wls_cap1] [WARNING] [] [oracle.ucp] [tid: [ACTIVE].ExecuteThread: '32' for queue: 'weblogic.kernel.Default (self-tuning)'] [userId: <anonymous>] [ecid: 6e4a19df-778e-4b08-b085-089041f8ceb3-00003564,0] [APP: Capps] [partition-name: DOMAIN] [tenant-name: GLOBAL] :::got unchecked exception from user-defined connection labeling callback

[2022-02-03T14:56:20.391-06:00] [wls_cap1] [WARNING] [] [oracle.ucp] [tid: [ACTIVE].ExecuteThread: '3' for queue: 'weblogic.kernel.Default (self-tuning)'] [userId: <anonymous>] [ecid: 6e4a19df-778e-4b08-b085-089041f8ceb3-00003565,0] [APP: Capps] [partition-name: DOMAIN] [tenant-name: GLOBAL] :::got unchecked exception from user-defined connection labeling callback

[2022-02-03T14:56:20.943-06:00] [wls_cap1] [WARNING] [] [oracle.ucp] [tid: [ACTIVE].ExecuteThread: '35' for queue: 'weblogic.kernel.Default (self-tuning)'] [userId: <anonymous>] [ecid: 6e4a19df-778e-4b08-b085-089041f8ceb3-0000356f,0] [APP: Capps] [partition-name: DOMAIN] [tenant-name: GLOBAL] :::got unchecked exception from user-defined connection labeling callback


Or it may not be related to ORDS upgrade and some other issue like runaway process. I will do more research, just wondering if anybody has noticed this with 21.4 ORDS?

«1

Answers

  • thatJeffSmith-Oracle
    thatJeffSmith-Oracle Distinguished Product Manager Posts: 8,510 Employee

    Do you see anything if you run ORDS as a standalone process?

  • Jim  Schmitz
    Jim Schmitz Member Posts: 15 Bronze Badge

    I really don't want to set it up standalone, unless I can do that without impacting the one running? I am trying to track down the weblogic logs against the apex process and see if it could be some runaway automated processes kicking off. It does not seem to correspond to any custom web services running. This could be a weblogic issue just a coincidence it happened after ORDS upgrade....

  • thatJeffSmith-Oracle
    thatJeffSmith-Oracle Distinguished Product Manager Posts: 8,510 Employee

    You can multiple mid tiers all hitting the same database, no problem.

    If ORDS is happy, then you can know for sure it's something to do with WLS.

    However, ORDS handles the connection pools itself, not WLS. So these [ucp] could be a bad omen.

  • Jim  Schmitz
    Jim Schmitz Member Posts: 15 Bronze Badge

    These messages are displayed whenever you execute an Apex page, even a refresh and the initial login page. The message seems to come before the page is rendered. The weird thing is, some of our really old applications, that were initially done in Apex 5 do not cause this error. The newer apps have changes in global page 0 and login page 101, but I deleted them as a test and the message still spits out in Weblogic log right before the "page not found" occurs. No javascript errors. I also made sure the app was updated with new theme. What kind of things should I check for? The connections I see in the DB seem to match the params in default.xml.

  • thatJeffSmith-Oracle
    thatJeffSmith-Oracle Distinguished Product Manager Posts: 8,510 Employee

    Can you share your version of WLS and Java? Our inhouse WLS 14 and Java 11 aren't showing these in the logs.

    Also, are you seeing these every few minutes, on every request, only at startup?

  • devegied2
    devegied2 Member Posts: 3 Red Ribbon

    I have deployment of APEX  21.2.1 + ORDS 21.4.0 on Apache Tomcat 8.5.73 and have same warnings in my logfiles. Also there is huge performance problem because all requests that have such warning logged takes at least 1200 milliseconds to complete (I logged it in Tomcat access log with millisecond precision). I created public page with pre-header process

    owa_util.mime_header('text/plain', TRUE);
    htp.p(to_char(localtimestamp,'HH24:MI:SS.FF3'));
    apex_application.stop_apex_engine;
    

    If I debug such page APEX debug log shows that execution from start to finish takes about 20 milliseconds but tomcat logs oracle.ucp warning and shows that request took from 1200 to 2800 milliseconds.

    After some more digging I found in audit trail that APEX_REST_PUBLIC_USER tries to do proxy connection as apex workspace user and fails (ORA-1017 in audit trail). Then it tries again and succeeds (with PROXY_USER=APEX_REST_PUBLIC_USER, AUTHENTICATED_IDENTITY=APEX_PUBLIC_USER, CURRENT_USER=workspace user, SESSION_USER=APEX_PUBLIC_USER).

    I rest enabled workspace and registered schema with ORDS. After these steps there is no warnings in the logs and tomcat logs that response was served in around 40 milliseconds. In audit trail I see that now ORDS_PUBLIC_USER is doing proxy connection to apex workspace user. Runtime environment also is a bit different (PROXY_USER=ORDS_PUBLIC_USER, AUTHENTICATED_IDENTITY=APEX_PUBLIC_USER, CURRENT_USER=workspace user, SESSION_USER=APEX_PUBLIC_USER).

    This apex instance was freshly installed, but all configuration was done using APEX_INSTANCE_ADMIN API. Workspace was exported from APEX 21.1.1 + ORDS 20.4.3 instance and imported by running export file in SQL*Plus. It could be that at time of import some step was missing.

  • Jim  Schmitz
    Jim Schmitz Member Posts: 15 Bronze Badge

    We are using <Version: WebLogic Server 12.2.1.3.0>. I just patched it with the latest January 2022 Critical Patch Update to see if that made a difference.

    OPatch Patches:

    33699205;24561064;Sat Feb 05 15:28:33 CST 2022;WLS PATCH SET UPDATE 12.2.1.3.211222

    33678607;24558585;Sat Feb 05 15:25:57 CST 2022;RDA release 20.4-20211126 for OFM SPB

    33290784;24421782;Sat Feb 05 15:25:35 CST 2022;JDBC 12.2.0.1 FOR CPUJAN2022 (WLS 12.2.1.3)

    1221317;24516525;Sat Feb 05 15:25:17 CST 2022;Bundle patch for Oracle Coherence Version 12.2.1.3.17

    java -version

    java version "1.8.0_251"

    Java(TM) SE Runtime Environment (build 1.8.0_251-b08)

    Java HotSpot(TM) 64-Bit Server VM (build 25.251-b08, mixed mode)

    We are not on the latest java 8, I can try updating that and see if it makes a difference.

    However, I did find another key. It only produces the message if I have "Friendly URLs" turned on. The older applications I referred to in earlier post, we never converted them to Friendly URLS. All new applications since we got to Apex 20, we always use Friendly URLs. If I turn off Friendly URLs, it goes away. If I turn it back on, it comes back. That is also why the Apex GUI itself does not cause the log warning, it is not using Friendly URLs.

  • Jim  Schmitz
    Jim Schmitz Member Posts: 15 Bronze Badge

    So further digging, I seem to have the same issue as devegied2. "1017" errors.

    However, I have checked everything and at a loss on where to "fix" this? The restful services are properly registered to "APPSC", in this case. I can logon with all the passwords, I even reset the "conf" xml files with the passwords in case that was out of whack. I ran the apex verify script, that looks good. Also the ords validate. This is an APEX upgrade and an ORDS 21.4 install. Both were ran as "SYS" user. Thoughts??

  • Jim  Schmitz
    Jim Schmitz Member Posts: 15 Bronze Badge
    edited Feb 8, 2022 3:54AM

    This looks like it might be the issue: https://docs.oracle.com/en/database/oracle/application-express/21.1/htmrn/index.html#HTMRN-GUID-B08BCCD3-20FD-4E9C-90B3-997A6CE5ED2A 9.1.8.3 Issue when Using SQL Developer or SQL Scripts to Enable ORDS on an Application Express 18.1 or later Schema. We started with Apex 5.1, and separately at some point created webservices using the webservice API (not thru APEX). The webservices and apex both used the same schema. However, the ORDS SCHEMA ALIAS and the APEX PATH PREFIX were different. I found this below where you can override the ORDS Schema, but the note implies you should really not do this. So I just did the resolution from that note, picked the Schema Alias (even though it was the same one), and clicked save and that seems to have reset everything. The Unchecked Error/1017 errors are no longer happening.  But now I have to remediate some code that has the Apex Path Prefix. But, I would rather change apex than the webservice since I have more control over that one.

    This option is so new, I am not getting any hits "Googling" it. I guess it is not recommended to have different alias's to the same schema?


  • User_VTSZE
    User_VTSZE Member Posts: 2 Green Ribbon

    Yes, it reproduces in ORDS Standalone. I see the warnings in the console. Apex is wroking fine.

    2022-02-09T12:49:11.527Z WARNING   *** jdbc.MaxLimit in configuration |apex|rt| is using a value of 10, this setting may not be sized adequately for a production environment ***

    2022-02-09T12:49:11.531Z WARNING   *** jdbc.InitialLimit in configuration |apex|rt| is using a value of 3, this setting may not be sized adequately for a production environment ***

    2022-02-09T12:49:14.376Z INFO    Oracle REST Data Services initialized

    Oracle REST Data Services version : 21.4.0.r3481956

    Oracle REST Data Services server info: jetty/9.4.44.v20210927

    2022-02-09T12:49:29.037Z WARNING   :::got unchecked exception from user-defined connection labeling callback

    2022-02-09T12:49:30.588Z WARNING   :::got unchecked exception from user-defined connection labeling callback

    How to get rid of this "got unchecked exception" warning

    Thanks in advance

    revion