Home | Contact Us | FAQ | Search & Site Map | Link to Us
Sign In | Join | Other 45 Sites in Network
Home
Discussion GroupsGeneralPHPASPPerlColdFusionFlashHTML, CSS, ScriptsBrowsers

Webmaster Forum / ASP / Database Access / June 2009



Tip: Looking for answers? Try searching our database.

NLB Cluster Intermittent SQL2K5 Connection Problems

Thread view: 
Enable EMail Alerts  Start New Thread
Thread rating: 
Mike Wazowski - 27 May 2009 20:14 GMT
Hi Guys

I know this is really awkward but for the last three days now at around 5:30
PM our two (NLB clustered) webservers have been unable to connect to our SQL
2005 SP2 server.  The SQL server installation is configured in a failover
cluster.  The webservers are running Windows 2003 Standard Edition SP2 with
MDAC 2.8 SP2 (each has 4 quad core Xeons and 3 Gb RAM).  The servers are
running classic ASP websites and both servers exhibit the common error
messages to the visitors:

DBNETLIB][ConnectionOpen (Connect()).]SQL Server does not exist or access
denied

DBNETLIB][ConnectionOpen (SECDoClientHandshake()).]SSL Security error

Rebooting both webservers resolves the problem.  Another webserver running a
copy of the clustered websites, which is not used by the public, is still
able to access all the websites/databases that link to the SQL Server.  We
don't think the problem is with the SQL Server set-up.  When we failed over
from one to the other the problem persisted.

No changes or tasks have been made/added for months.  Things have been
running smoothly for a while.

Nothing is logged in the event log on the webservers or the logs within SQL
Server.  I don't know where to go next.

Does anyone have any advice?

Many thanks

Mike
Mike Wazowski - 28 May 2009 20:37 GMT
Happened again tonight at 7:10 PM.  This time I only rebooted one of the
cluster servers and it did the trick.  It must be something to do with NLB
or sockets or something...

I don't know where to go.  Nothing is logged anywhere.

> Hi Guys
>
[quoted text clipped - 28 lines]
>
> Mike
Mike Wazowski - 03 Jun 2009 14:04 GMT
Logged a call with Microsoft.  After some Netstat analysis they have
suggested increasing the number of ports (MaxUserPorts) and the
TcpTimedWaitDelay registry settings on the two NLB servers.  Apparently
they're nearing the default of 5000 so possibly out of ports.

> Happened again tonight at 7:10 PM.  This time I only rebooted one of the
> cluster servers and it did the trick.  It must be something to do with NLB
[quoted text clipped - 34 lines]
>>
>> Mike
Bob Barrows - 03 Jun 2009 15:08 GMT
Good to know - thanks for the update.

It will be interesting to hear if this works for you. I've not run into
the problem myself, but I'm sure it could crop up someday.

> Logged a call with Microsoft.  After some Netstat analysis they have
> suggested increasing the number of ports (MaxUserPorts) and the
[quoted text clipped - 43 lines]
>>>
>>> Mike

Signature

HTH,
Bob Barrows

Mike Wazowski - 08 Jun 2009 19:54 GMT
Yes, well it's Monday and we're past the previous times of failure so
fingers crossed.  The two webservers in the NLB cluster have been putting
out between 35 and 40 Mbps each for most of the day so it's looking good so
far.  Incoming is about 10Mbps but of course that's duplicated to both
servers.

> Good to know - thanks for the update.
>
[quoted text clipped - 48 lines]
>>>>
>>>> Mike
 
Sign In
Join
My Latest Posts
My Monitored Threads
My Blog
My Photo Gallery
My Profile
My Homepage

Start New Thread
Enable EMail Alerts
Rate this Thread



©2010 Advenet LLC   Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.