Server Port Not Connecting, but SDE Service is OK?

[847396730]

Hello!  Our production SDE geodatabases (SQL Server 2008 R2) are accessible via the SDE service, but suddenly unavailable via the SQL server port (direct connect). This means I cannot access the geodatabase in SQL Server Managment Studio, which is quite a handicap.  This also impacts our 10.1 and 10.2 clients (though most production users are currently on 10.0).

All my attempts at direct connect fail with a timeout and/or general failure to access dbms server.

How can I troubleshoot this?

Thank you!

ArcGIS Desktop 10.0, 10.1, 10.2
ArcSDE for Microsoft SQL Server 64 bit v10 SP3

[VinceAngelo]

Can you use a SQL-Server client like 'sqlcmd' to access the database?  It sounds like
something has happened to prevent the database from accepting connections (which
doesn't really have anything to do with ArcSDE).

How long has it been since your last database restart?

Have you done any software package updates on the server?

Any firewall configuration?

- V

[AsrujitSengupta]

Just like Vince suggested, could be a Firewall security issue.

Try turning off the Firewall for a test connection. If the connection goes through after turning off the Firewall, you might need to add the ports that SQL Server uses(1433 & 1434) to the Firewall exceptions.

Regards,

[847396730]

Thank you both, for the replies.  I can't access the server directly; our IT Division does not allow it.  Sadly, it appears I may need to wait for them to get back to me.  I am not aware of any changes having been made to the firewall setup, or recent software package updates, and I know for certain this problem started very recently.  So while the firewall is unlikely to be the culprit, I cannot perform the tests to exclude it.

sqlcmd -s yields a timeout error.

I will follow up when I have more to offer.  Thanks again.

[VinceAngelo]

If 'sqlcmd' fails, the issue is the database instance (be it failure or firewall), not anything
within the Esri stack.

I'd think this would be a show-stopper priority ticket, not a "when they get back to me"
thing.  At sites where I work the IT department loses money when a Pri-1 ticket isn't
resolved within an hour, and IT managers can lose their jobs if the outage causes
award fee degradation due to a missed availability metric.

- V

[847396730]

Thanks, Vince.  I have a meeting this morning with our IT folks to take a look.  Like so many others in this industry, our relationship with IT is delicate, and since our users aren't down (they're on version 10, using the SDE service, not direct connect), I chose not to "correct" IT's assignment of "Priority - Standard" on this one.  But I did catch someone in the hall, who understood the problem's magnitude, so hopefully I will have the solution soon. 

Thanks again for the direction; these tips plus the Google-learn have been informative!

[VinceAngelo]

The most likely cause is that they turned on a firewall, not realizing you needed
direct access to the database.

- V

[847396730]

Hello!  To follow up on this thread: we never found the cause.  When I arrived for the meeting, they had already bounced the database (unfortunately causing some lost edits), which cleared the problem.  Their description of what they saw was too vague to pin down, so we'll have to wait until it happens again to learn more. Thanks again for the input!