Frequent General Network error

Question

Post reply

Frequent General Network error

BOVA Systems

Valued Member

Points: 71
More actions
June 29, 2010 at 6:41 am

#234434

We installed a new server with only SQL Server 2008 on it in Plant A.
In Plant A 20 computers are connected via LAN. Here everything works fine.
In Plant B 10 computers are connected via WAN. Each user gets a 'General Network Error' at least 10 times a day.
Is there any parameter I could change to avoid these problemes ?
Application is written in VB6. Database used to be SQL2000 but we attached the 2000 - databases to SQL2008 when we installed the new server.
Please help me.

Viewing 15 posts - 1 through 15 (of 15 total)

You must be logged in to reply to this topic. Login to reply

smitty-1088185 SSCrazy Points: 2321 More actions · Answer 1

How often do the errors occur, once or twice a day or every 15 minutes?

Is everyone in the plant getting the errors or a selected few?

What time of day does the errors occur, same time every day?

Jason Messersmith Right there with Babe Points: 794 More actions · Answer 2

Jason Messersmith

Right there with Babe

Points: 794

June 29, 2010 at 9:23 am

#1186460

Server 2003 or 2008?

BOVA Systems Valued Member Points: 71 More actions · Answer 3

Hi,

Everyone is getting the errors when they occur. I think it is because of some parts of the program that is consuming a lot of 'database' or 'network' time.

Yesterday I thought of it that maybe the crystal reports (written in v 8.5) may cause the problem.

On the server Win Server 2008 is running

smitty-1088185 SSCrazy Points: 2321 More actions · Answer 4

The first couple steps is process-of-elimination.

Run the following command from a workstation that is getting the errors.

C:\>ping -t MYSQLSERVER>c:\monitoring.log

Let the ping process run all day and make sure you have enough drive space

for the ping log. You can view the log while its running just say no when it ask

you to save it.

If you see high response times from the reply during the timeout errors you have

a network connectivity issue. Try updating the NIC card firmware and upgrade

the drivers also. Use a freeware program like Wireshark and analyze the packets.

If the response times are all under 100 you probably have an application problem.

Run "sp_who2" on the SQL server from the SQL query analyzer and see there is any

blocking going on.

Create an ODBC connection from the workstation and click the connection test while

the errors are occurring.

BOVA Systems Valued Member Points: 71 More actions · Answer 5

Good morning,

I put the ping command as you advised.

Mostly we have a response time of 31 ms on the first PC and 46 ms on the second one. When the error occurs there are peeks of 600, 1900 even 2500 ms. Right after the 600, 1900 and 2500 ms everything turns back to 31 or 46 ms.

Marcos Leandro Rosa Mr or Mrs. 500 Points: 512 More actions · Answer 6

Flush DNS Cache and then try to access without doing anything else. Sometimes for dual homed networks

SQL gets timeout. The problem can be resolved by flushing DNS cache.

Regards.

Best Regards,
Marcos Rosa / marcosfac@gmail.com

BOVA Systems Valued Member Points: 71 More actions · Answer 7

Thx for the hint, but can you tell me how to flush the DNS Cache ?

Regards

BOVA Systems Valued Member Points: 71 More actions · Answer 8

Here are the results of the ping -t command. NOTHING was running on the workstation.

Can someone explain to me what could be the problem that causes the time out or high response times.

Probably something on the server I suppose ?

Antwoord van 192.168.0.70: bytes=32 tijd=31 ms TTL=126

Antwoord van 192.168.0.70: bytes=32 tijd=684 ms TTL=126

Antwoord van 192.168.0.70: bytes=32 tijd=2121 ms TTL=126

Time-out bij opdracht.

Antwoord van 192.168.0.70: bytes=32 tijd=304 ms TTL=126

Antwoord van 192.168.0.70: bytes=32 tijd=355 ms TTL=126

Antwoord van 192.168.0.70: bytes=32 tijd=31 ms TTL=126

Regards

Marcos Leandro Rosa Mr or Mrs. 500 Points: 512 More actions · Answer 9

To flush DNS Cache you use:

On CMD:

ipconfig /flushdns

Best Regards,
Marcos Rosa / marcosfac@gmail.com

Stefan_G SSCertifiable Points: 6609 More actions · Answer 10

BOVA Systems (7/1/2010)
Here are the results of the ping -t command. NOTHING was running on the workstation.
Can someone explain to me what could be the problem that causes the time out or high response times.
Probably something on the server I suppose ?
Antwoord van 192.168.0.70: bytes=32 tijd=31 ms TTL=126
Antwoord van 192.168.0.70: bytes=32 tijd=31 ms TTL=126
Antwoord van 192.168.0.70: bytes=32 tijd=31 ms TTL=126
Antwoord van 192.168.0.70: bytes=32 tijd=31 ms TTL=126
Antwoord van 192.168.0.70: bytes=32 tijd=31 ms TTL=126
Antwoord van 192.168.0.70: bytes=32 tijd=684 ms TTL=126
Antwoord van 192.168.0.70: bytes=32 tijd=2121 ms TTL=126
Time-out bij opdracht.
Antwoord van 192.168.0.70: bytes=32 tijd=304 ms TTL=126
Antwoord van 192.168.0.70: bytes=32 tijd=355 ms TTL=126
Antwoord van 192.168.0.70: bytes=32 tijd=31 ms TTL=126
Antwoord van 192.168.0.70: bytes=32 tijd=31 ms TTL=126
Antwoord van 192.168.0.70: bytes=32 tijd=31 ms TTL=126
Antwoord van 192.168.0.70: bytes=32 tijd=31 ms TTL=126
Antwoord van 192.168.0.70: bytes=32 tijd=31 ms TTL=126

This sounds like a low-level network issue. It could be a malfunctioning network card somewhere, an overloaded network segment, or some kind of configuration problem.

You need help from a network expert who can set up monitoring of the network to see what is really happening.

This is NOT an application problem and it is NOT a SQL server problem.

smitty-1088185 SSCrazy Points: 2321 More actions · Answer 11

You have a networking issue like the others have stated. Use two different workstations and ping two different servers at the same time. This will tell you if its a server issue or Cisco switch problem.

From PlantB Workstation1

Ping -t SQL_SVR01

From PlantB Workstation2

Ping -t SQL_SVR02

If you get high response times from both workstations you probably have a transport problem

with one of your switches or routers.

Run a Trace Route command from both workstations like this:

C:>\tracert SQL_SVR02

Look for high response times from one of the network hops. Post the results when you are done.

I am more than welling to help you but you may want to post the results on a networking forum since its not a SQL issue.

Rick.Cornell.SME SSCrazy Eights Points: 8597 More actions · Answer 12

Some things to look at also:

1. What protocols are you using, I tend to shut everything down except TCPIP.

2. Make sure your Service Principal Name is set correctly, ask you Domain Admin to review settings.

3. Make sure your DNS settings are correct in Active Directory again work with your Domain Admin to review settings.

4. It always good to update adapter drivers on server and\or clients.

5. Do you get the same results when you ping by Name or IP across the WAN?

BOVA Systems Valued Member Points: 71 More actions · Answer 13

I did a tracert and that is the result I became :

Tracing route to server1 [192.168.0.70]

over a maximum of 30 hops:

1 <1 ms <1 ms <1 ms 10.205.190.253

2 23 ms 13 ms 21 ms 192.168.201.1

3 24 ms 22 ms 15 ms server1 [192.168.0.70]

Trace complete.

smitty-1088185 SSCrazy Points: 2321 More actions · Answer 14

Those results look normal but you'll have to run the tracert when you are getting the time-out errors. Are you sure that was test across the WAN from PlantA to PlantB? There's only one internet hop in those results.

For security purposes do not post the actual IP addresses, just the response times.