Expand my Community achievements bar.

SOLVED

Authentication fails after saving OSGI config

Avatar

Level 2

I've been investigating how to set a password validation string but in the process have a very unusual behaviour.  I am using AEM 6.2.

Starting with a freshly-installed system at http://localhost:4502/system/console/configMgr, I search for 'Apache Jackrabbit Oak AuthorizableActionProvider', click on that name to open the configuration to open the configuration dialog, then click 'Save' from the buttons in bottom-right corner of the dialog.  Do not make any configuration changes, just open the dialog and click 'Save'.

There is a message at the top of the window after this: 'Configuration Admin Service is running.'

Unfortunately, CRXDE is now broken at: http://localhost:4502/crx/de/index.jsp​.  The error message is:

Also, I cannot login to the author at: http://localhost:4502/​  The error message is:

The java process that is running the author instance has to be force-shutdown.  After a restart, everything is (apparently) working again.

I am concerned that this could cause issues on deployment to production.

What should I be looking for to analyse and resolve this?

1 Accepted Solution

Avatar

Correct answer by
Level 5

johng23

There are some known issues in AEM 6.2 GA instance. You can see issues notes in 6.2 SP1 release notes such as

  • Changing Oak-related OSGI configuration makes the AEM instance dysfunctional. GRANITE-11906
  • Fixed issues around instances becoming non-responsive when Sling Resource Resolver Factory is updated. GRANITE-13126

I suggest you update your instance to the latest SP+CFP which would be SP1 + CFP14 and re-try to see if you face any issues. I've just performed a quick test on mine SP1+CFP14 instance and didn't face any issues.

Regards,

Aneet

View solution in original post

8 Replies

Avatar

Correct answer by
Level 5

johng23

There are some known issues in AEM 6.2 GA instance. You can see issues notes in 6.2 SP1 release notes such as

  • Changing Oak-related OSGI configuration makes the AEM instance dysfunctional. GRANITE-11906
  • Fixed issues around instances becoming non-responsive when Sling Resource Resolver Factory is updated. GRANITE-13126

I suggest you update your instance to the latest SP+CFP which would be SP1 + CFP14 and re-try to see if you face any issues. I've just performed a quick test on mine SP1+CFP14 instance and didn't face any issues.

Regards,

Aneet

Avatar

Employee

I can't replicate this issue on latest maintenance release of AEM 6.2. I would suggest first update to latest 6.2 patch release (1) and let me know if things look good.

(1) : Adobe Experience Manager Help | Release Notes: AEM 6.2 Cumulative Fix Pack

Avatar

Employee Advisor

Hi,

when you re-configure Oak services, typically a lot of AEM is restarted; that's because this service restarts, but before that OSGI shuts down all depending services, the services these depending services depend an so on. In most cases that's comparable to a complete restart of AEM.

When testing this, you should validate in your log file when the services restart is over, and then everything should be fine again.

Jörg

Avatar

Level 2

Thanks Aneet,

I've noticed that CFP15 is available (since 6 June).  Is there any reason I should still take CFP14 instead?

John

Avatar

Level 6

Hi Johng23,

As General paractice , for all instance which we start need to ensure that all bundles are up and running fine .

Once we found this details on OSGI console we are good to navigate to other consoles.

I guess in your case you made changes while some bundles are not yet ready and so you faced issue on opening crx/de console.

Avatar

Level 5

Hello johng23

We should aim to install latest CFP available for the version so yes you can surely install CFP15.

Best Regards,

Aneet

Avatar

Level 2

These are my findings from trying out the suggestions.  This was my first upgrade so I read the documentation carefully and did a number of tests.  Since SP1 requires Hotfix 12785, and CFP15 is optional after SP1, I tried the following combinations:

  • HF12785 on its own: did not fix the problem (as expected)
  • SP1 (including HF12785): fixed the problem
  • SP1+CFP15: fixed the problem

All the answers above were helpful but I have chosen aneeta45259594's detailed answer as the correct one.  Thank you!