Solved

Intermittent execution issues

  • 23 May 2023
  • 5 replies
  • 57 views

Userlevel 1
Badge +1

I’m regularly running into the “Cannot open server 'sql-instances-prod' requested by the login. Client with IP address '125.7.127.171' is not allowed to access the server.” error message, especially during overnight executions by the execution scheduler.

But I can’t replicate the connection issue, i.e. it seems to be intermittent:

Has anybody else encountered this issue and do you know what would be causing it / what the workaround might be?

icon

Best answer by Søren Sørensen 26 May 2023, 14:06

View original

5 replies

Userlevel 5
Badge +5

Hi @Søren Sørensen Microsoft experienced some outages on Azure SQL DBs last week which may be the cause of the issue you are seeing. This should have been resolved by Friday. Has the error occurred since then?

Userlevel 1
Badge +1

Hi @Christian Hauggaard 

Yes, I can unfortunately replicate it. I’ll check with our customer … maybe there’s something wrong in their setup. I might also reduce the concurrency to see if that helps.

Userlevel 5
Badge +5

Hi @Søren Sørensen Did you experience the same issue during the nightly scheduled execution the weekend or last night? Is this what you mean when you say you can replicate it, or can you replicate it outside of the scheduled nightly load? If you could please send screenshots of the execution log that would be much appreciated.

Userlevel 1
Badge +1

Hi @Christian Hauggaard

I had some timeout issues, which could or could not be related, but I haven’t had this thread’s problem since I changed the ODX timeouts. Let me monitor things for another week … if the issue pops up again, I’ll get back to you and pick your brain.

PS. I can only see the error messages in the Event Log; TX reports failure but no detailed messages.

Userlevel 1
Badge +1

Hi @Christian Hauggaard 

Microsoft agrees that there are many intermittent errors in Azure and that the best solution is an abundance of retries. The last couple of days’ testing support this … my takeaway is that I should

  1. increase timeouts in general and data source timeouts in particular and
  2. have large numbers of retries in place in MDW.

These two simple rules have further stabilized ODX and MDW executions … fingers crossed.

Reply