Natural getting NAT3148 error from time to time

, , ,
1

Does anyone know is there any setting or configuration of adabas session or database restart? We are getting NAT3148 error from Natural in random time and I can see %ADANUC-I-TERMINATED and %ADANUC-I-DBSTART message in adanuc.log. The logs generated when Application getting NAT3148 error. Seems like this happening in random time but frequently. We do not have any scheduled task/cron job to restart ADABAS.

Any help or insight would be appreciated.

2

I am not sure how better to respond than to say check with your DBA as to whether your database is being brought down and up at “random times”, which I think precludes the idea it is scripted. Either the DBA is deliberately bouncing the database for some reason or the database is in need of some care if it crashes (which doesn’t happen for a properly managed ADABAS on any platform). I hope this is at least a non-production environment.

3

I agree with Brian - there are two considerations here: 1) something is bringing your database down. If it is a crash, there should be information in the adanuc.log as to what the crash is. If the termination is “normal”, then there is something/someone initiating a shutdown using Adabas utilities. 2) you most likely have some Linux function that is monitoring the status of the Adabas process and initiating a restart if it is down. Work with your Linux systems administration team to find out how this is done (systemctl? nagios? ??)

4

Thank you for the responses. Actually I’m the DBAs and investigating why the ADABAS is going down in random time. @Douglas_Kelly, is there any chance that ADABAS can be crushed when there is no enough resource available on the host then it terminated and auto start again? I do not see any error in adanuc.log but termination then start again.

5

Check if there are any files named

ADABAS.<dbid>.<timestamp>
SAGSMP.<dbid>.<timestamp>

in your database directory.

These are generated when the database crashes,

If there are any

  • check if the timestamps relate to the times when you receive the Rsp148.
  • open a support incident and attach one or more of these files to the incident for investigation
6

Thank you Wolfgang. Unfortunately, I do not see the crash files you mentioned in above. adanuc.log. and NUCPLG.xxxx files.

7

Look in your DBnnn.INI file NODE_PARAMETER section ANALYSER - if ACTION=YES, then AEO is activated and the actions in ADANUC_TERMINATED will be used to restart the database automatically.

Are there messages indicating that the database is terminated normally, cancelled or aborted?

If the database is not crashing, then there is likely some action that is stopping the adanuc process. Check with your Linux systems team to find what is terminating it - adastop? adakill? adaend? kill? There may be information in files like /var/log/messages or /var/log/secure around the time the database comes down.

8

Thanks Kelly. I agree that there should be a process/task running to restart the database. I will keep digging around. Thank you for your responses!