gbarnes0990
asked on
Slow network drives
Hello,
We have just built a new file server and some of the users are complaining that access to the shared drives is slow.
The server is a DL380 with dual 2.8ghz processors. When the users go into Word or Excel and try and access their files the process hangs when going through the directories to get to the file.
The server has 2 x 1gb nic cards and these are both set to 1000/full they are also both teamed. The 2 network cables coming out of the server are going to 2 seperate cisco switches. Both ports on the switches have also been set to 1000/full. Both the server and the cisco switch were originally set to auto/auto (server) and a-full-1000 (switch) we have changed to the current setup to see if there is any improvement.
We have just built a new file server and some of the users are complaining that access to the shared drives is slow.
The server is a DL380 with dual 2.8ghz processors. When the users go into Word or Excel and try and access their files the process hangs when going through the directories to get to the file.
The server has 2 x 1gb nic cards and these are both set to 1000/full they are also both teamed. The 2 network cables coming out of the server are going to 2 seperate cisco switches. Both ports on the switches have also been set to 1000/full. Both the server and the cisco switch were originally set to auto/auto (server) and a-full-1000 (switch) we have changed to the current setup to see if there is any improvement.
Also check the client machines duplex settings on the NIC. They would need to match the 1000/Full through every switch back to the server, which could be a few different switches depending your environment. If the client machine only has a 10/100 card, then it's probably best to double-check that all of the ports are set to auto.
Have there been any changes with DNS? How are you mapping to the network shares? If it is a new file server, some of the users may have a shortcut or map to the "old" server and this would cause a major delay/hangup.
I think I would start by keping it simple. A file server should be plenty fast with just a simple ethernet connection, so I advise running it on just one NIC card and see what happens. I would also set the switches back to default auto settings.
What type of teaming is being used?
Review the HP network teaming configuration utility for errors under status.
Type of switches (Cisco)?
Review the HP network teaming configuration utility for errors under status.
Type of switches (Cisco)?
What type of teaming is being used?
Review the HP network teaming configuration utility for errors under status.
Type of switches (Cisco)?
Review the HP network teaming configuration utility for errors under status.
Type of switches (Cisco)?
Are there any errors in any of the 3 event logs ?
DOes perfmon show any bottlenecks ?
I hope this helps !
DOes perfmon show any bottlenecks ?
I hope this helps !
ASKER
Cheers for the advice guys.....
All the client machines should have gigabit nic cards. I am getting them changed now from auto to 1000/full. There have been no changes to DNS. The users get something like 3 mapped drives the login script was amended so that the old server name was replaced with the new. None of the users should be using a shortcut to the old server (which has been switched off). Teaming has been setup through HP Network Config Utility. There are only 2 on board cards these have both been teamed. The switches are Cisco 35xx and there are no errors or dropped packets.
I have checked HP SIM and there are no errors I have run extended pings to the switches and all pings are 1ms with no dropped packets. Event viewer shows no errors. The HP Proliant Integrated Log shows no errors.
I will run perfmon to see if this sees any problems.
Cheers,
All the client machines should have gigabit nic cards. I am getting them changed now from auto to 1000/full. There have been no changes to DNS. The users get something like 3 mapped drives the login script was amended so that the old server name was replaced with the new. None of the users should be using a shortcut to the old server (which has been switched off). Teaming has been setup through HP Network Config Utility. There are only 2 on board cards these have both been teamed. The switches are Cisco 35xx and there are no errors or dropped packets.
I have checked HP SIM and there are no errors I have run extended pings to the switches and all pings are 1ms with no dropped packets. Event viewer shows no errors. The HP Proliant Integrated Log shows no errors.
I will run perfmon to see if this sees any problems.
Cheers,
ASKER
Perfmon is showing no bottlenecks. The users PC's are set to auto they can't be set to 1000/full as they are all connected to 10/100 switches. The old file server was an IBM xseries 240 with 2 teamed nic cards which were both 10/100. We never had speed issues with this server??
Here are a couple of MS related docs that talk about a problem with SP2 on their Windows 2003 servers. I didn't think they were having a problem but evidentally some servers are showing a slow response to network shares after the upgrade.
http://support.microsoft.com/kb/936594
http://support.microsoft.com/kb/916846
http://blogs.technet.com/sbs/archive/2007/04/24/common-networking-issues-after-applying-windows-server-2003-sp2-on-sbs.aspx
http://support.microsoft.com/kb/936594
http://support.microsoft.com/kb/916846
http://blogs.technet.com/sbs/archive/2007/04/24/common-networking-issues-after-applying-windows-server-2003-sp2-on-sbs.aspx
What type of teaming?
ASKER
We have had problems with SP2 as well which is why we haven't installed it.
Teaming is setup as follow:-
TEAM TYPE SELECTION = AUTOMATIC (RECOMMENDED)
TRANSMIT LOAD BALANCING METHOD = AUTOMATIC (RECOMMENDED)
Teaming is setup as follow:-
TEAM TYPE SELECTION = AUTOMATIC (RECOMMENDED)
TRANSMIT LOAD BALANCING METHOD = AUTOMATIC (RECOMMENDED)
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
We got burnt not using fail on fault. Good recommendation to go back to a single NIC and work forward.
How was the Fail on fault configured when the burning happened, was it set to manual?
Also see the page 49 from the link listed below: http://h71028.www7.hp.com/ERC/downloads/4AA0-1139ENW.pdf?jumpid=reg_R1002_USEN
Also see the page 49 from the link listed below: http://h71028.www7.hp.com/ERC/downloads/4AA0-1139ENW.pdf?jumpid=reg_R1002_USEN
ASKER
No routers involved. Going to give NFT ago but won't be able to do this till sunday night. Cheers for all your help guys.
No, we got burnt NOT using fail on fault. The team was configured for some kind of load balancing controlled by the switch (by the former net admin) and it caused numerous errors.