NAS (nassql.exe) stops working when applying hotfix 36259

rocatis

I have a NAS service (build 32012) which works perfectly on a Windows 2008 R2 server.

When I install hotfix 36259 (by simply copying the hotfix files into my NAS folder), I can no longer start my NAS service. I get an error saying "Error 1053: The service did not respond to the start or control request in a timely fashion". Nothing is posted in the Event Viewer.

I can run the NAS in a DOS prompt. I have two other servers (Windows 2003) where the hotfix has been installed. The servers are virtually identical (one is the production server, the other is the test server) but the hotfix only works on one of them.

I've tried copying the NAS files from the working (test) server to no avail. As we did not have any problems installing the hotfix on the test server, it was installed on the production server - where it doesn't work. As the production database has been converted by the hotfix, rollback is not an option.

Any ideas?

dave_c

I installed a hotfix a while ago which did the same thing when I tried starting the middle tier service. I opened a case with MS and it turned out it required an internet connection to update some files the first time you ran it after applying the hotfix. Does you production server have an internet connection?

rocatis

That sounds weird...

The production server does not have internet access, and the chances that IT is going to allow this are slim to none.

einsTeIn.NET

It could be a combination of that Build No. and a certain Windows Hotfix package. Maybe that Windows package is rolled out to just one of the servers. This package leads to a system behaviour, where Windows wants to check on every start of a Microsoft service if there's an update available. So, it needs to connect to the internet to check if there's an update of that service abvailable.

We installled this package by regular system update. You won't notice that there's an issue until you restart your service/machine. And that's what it makes that hard to identify.

Since this is a Windows issue, you could have that issue with NAS as well as one of the Middle Tier Services.

We entered the right proxy in the internet options of the NAS service user to grant access to the internet. But I fully understand if there are some concerns about that solution.

einsTeIn.NET

Please also have a look at this post http://www.mibuso.com/forum/viewtopic.php?p=275392#p275392.

rocatis

It looks like I've fixed the problem. First of all, this KB article suggests to increase the timeout value from 30s (default) to 60s: http://support.microsoft.com/kb/922918
I did that, but at first it didn't work. It seems that you need to restart the server before this registry setting is activated which took me quite a while to realize.

That allowed me to start the services manually, but even though the services were set to start automatically, they didn't do that. The Event Log was empty.

It turns out that the 1053 timeout error does not put anything in the Event Log which means that timeout errors when rebooting the server really cannot be diagnosed (when you realize the services have not been started automatically, you tend to start them manually - which does not give you an error).

I set the "Startup Type" for the services to "Automatic (Delayed Start)" which solved that part of the problem, so right now everything seems to work just fine.

I'm slightly worried that a relatively simple hotfix installation can have such a profound impact on a system which otherwise runs fine.

Oh, and also: the build number shown in the Event Log (32012) does not match the build number of the nassql.exe file itself (36259). That didn't help the debugging process either...