1645Y
asked on
Server losing network connection randomly - reboot required to reactivate.
Ok, we are having an issue with one of our new servers.
It is a DL380 G5 - 1 quad core CPU with 4GB memory running Windows Sever 2003 SP2. We built this machine specifically for one application within our company, which required SQL server 2005. The server has 2 NICs which we connected to 1 subnet - teamed.
After I had installed WinServer2k3 + sp2 we left the machine sitting there for a month or so and we had no problems - then sql 2005 sp2 was installed and again - 1 month with no problems.
Since then, we have installed the application it was designed to support which involves an amount of data being imported into the SQL database. Either during the actual import of data, the SQL verification or the SQL DB backup (as far as we can tell) we lose network connection to the machine completely and I have to connect via iLo2 to reboot to fix. This will generally occur once every day or so.
When I connect to the machine via iLo the nic appeared connected, but can not contact anything on the domain outside its own ip (10.22.20.x & 127.0.0.1). I tried restarting the Network service - but this fails.
To diagnose this, we have changed the IP, port on the router, network cables. We then broke the team, disabled one nic and ran off the other - the same occured - we reversed that situation and the issue still occured. We had HP replace the motherboard as the NICs are onboard and the issue still occurs. There is nothing in the event log at all or anything in the SQL Server log that points to anything other than the network connection dropped a number of users connections. Have also upgraded software nic drivers/firmware with no results as yet and really starting to run out of ideas.
It is a DL380 G5 - 1 quad core CPU with 4GB memory running Windows Sever 2003 SP2. We built this machine specifically for one application within our company, which required SQL server 2005. The server has 2 NICs which we connected to 1 subnet - teamed.
After I had installed WinServer2k3 + sp2 we left the machine sitting there for a month or so and we had no problems - then sql 2005 sp2 was installed and again - 1 month with no problems.
Since then, we have installed the application it was designed to support which involves an amount of data being imported into the SQL database. Either during the actual import of data, the SQL verification or the SQL DB backup (as far as we can tell) we lose network connection to the machine completely and I have to connect via iLo2 to reboot to fix. This will generally occur once every day or so.
When I connect to the machine via iLo the nic appeared connected, but can not contact anything on the domain outside its own ip (10.22.20.x & 127.0.0.1). I tried restarting the Network service - but this fails.
To diagnose this, we have changed the IP, port on the router, network cables. We then broke the team, disabled one nic and ran off the other - the same occured - we reversed that situation and the issue still occured. We had HP replace the motherboard as the NICs are onboard and the issue still occurs. There is nothing in the event log at all or anything in the SQL Server log that points to anything other than the network connection dropped a number of users connections. Have also upgraded software nic drivers/firmware with no results as yet and really starting to run out of ideas.
Have you tried to reset tcp/ip stack?
What firewalls do you have installed on the server?
Can you ping the server when the card stops communicating?
Where can/cant you ping when it stops communicating?
Do you have physical access to the server when its down?
Can you ping the server when the card stops communicating?
Where can/cant you ping when it stops communicating?
Do you have physical access to the server when its down?
ASKER
No internal firewalls.
No, you can't ping the server from anywhere outside the server itself.
The only place you can actually ping its address is from the server itself.
When on the server you can not ping anything outside. Yes, we have physical access to the server when its down.
There does seem to be network activity lights on the machine when its down.
No, you can't ping the server from anywhere outside the server itself.
The only place you can actually ping its address is from the server itself.
When on the server you can not ping anything outside. Yes, we have physical access to the server when its down.
There does seem to be network activity lights on the machine when its down.
ASKER
Can't identify the exact cause, but have found something that I am fairly sure is the cause.
We had a contractor come in to do some DB copy from our old accounting system to new system in which he needed local admin access. Seems he installed Citrix Secure Access Client to connect to his work network from the server itself.
He then used his laptop / wyse box we gave him to connect to our server.
From what I understand there are issues with network compatibility that could cause a clash and bring down the specific IP. Trying to find definitive proof so I can go to the company with the data, but other than the fact I know he installed the software on the date the errors started to happen, and hasn't occurred since he left I can't find the exact cause.
We had a contractor come in to do some DB copy from our old accounting system to new system in which he needed local admin access. Seems he installed Citrix Secure Access Client to connect to his work network from the server itself.
He then used his laptop / wyse box we gave him to connect to our server.
From what I understand there are issues with network compatibility that could cause a clash and bring down the specific IP. Trying to find definitive proof so I can go to the company with the data, but other than the fact I know he installed the software on the date the errors started to happen, and hasn't occurred since he left I can't find the exact cause.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.