Network lagggg

I have been trying to figure this out for a while and I am not sure where to look for the answer.

I have a number of very powerful and very fast machines in our office. They run fine for the most part, but one aspect is really getting on my nerves.

When I connect to our office share drives on the network it seems to take quite a long time for the folder to load up. It seems to sit at a blank page until it finally populates it with all the files/folders and then everything seems to work fine and quickly, but then all of the sudden it get very slow and lagggs a lot when going from folder to folder.

The strange thing is I never get this type of problem on my computer.

What are somethings I should look for which can really hang up network drive access? Could it be related to the windows firewall or possibly an anti-virus? Or has anyone seen this type of lag before?

Hmmmm...........

1. Is your computer a part of this network?
2. By "office share drives" I assume that you mean file servers? is this correct?
3. How full are these drives? as in used space/free space.......once you fall below 20% free space you can expect performance issues.
4. When did you last defragment these shared drives?
5. Are you running Norton/Symantec?
6. Is your AV set to scan on open file and close file?
7. Is your AV set to scan in the background?
8. Do you have any automatic update processes enabled?
9. Is your AV set to scan heuristically?
10. When did you last run checkdisk?
11. Do all the machines use the same router/switch?
12. Do they all use the same cabling?

Just a few to be going on with :)

Yeah, this is a major problem for us. You name it it is server side. User documents? on the server. Common documents? on the server. Wallpaper? on the server. Desktop icons? on the server.

My top 2 guesses are:
1) both your client and the server are virus scanning the file when you open the folder
2) There is a bottleneck somewhere/too much traffic.

For us the problem is that our single 320Mbit raid can't keep up with 100 users. Disk access is 100% flat out, network is barely touched.

The best thing to do is load up windows performance monitor and put the big 4 on there.
1) % CPU time
2) % disk time
3) available memory
4) network output

then just spot which line hits the roof/floor when you open the folder. Since you refer to the machines as powerful my top two guesses are going to be hard drive hang up from either AV or users or network hangup from users or old infrastructure.

Quote from the future:

Quote:

---

Why don't you upgrade?

---

Theres no money.

Couple of things.. Speed/duplex settings on the server/workstation/switch (don't use auto/auto!).. And DNS not set up properly..

The server is a beast with 10K raid 5 sas drives. Dual dual core Xeon and 4GB ram with a 1Gb nic. It can't be the server and I would doubt its the client machines they are all top of the line and I always remove the norton or mcafee programs and usually put in AVG instead.

Brand new from Dell, but I never notice lag on my machine...then again I am on the same switch (Gbe switch) and a lot of people are on 100Mb switches.

I will try out the AV disabling or changing the settings to see if that has an affect. What should I set our switches to in order to get the maximum throughput? Would an upgrade to all Gbe switches do the trick?

I'll keep looking into this to see what is up and any information you care to drop on me is greatly appreciated.

An interesting test would be to attach your machine to one of the connections that displays lag and put that machine on your connection?

If yours doesn't lag and the other machine continues to then it is something on the desktop. Otherwise you need to look at the server and switch side.

Nihils suggestion would be able to rule out bad cable switch or port

could be a simple as that

Recently I had a problem wiith an onboard NIC ....sitting right beside a usb port...hooked up to an external drive. I would get frequent disconnects. threw in a pci nic...and the problem was solved. Appears the usb was causing interference with the onboard nic (never liked them realtecs anyway)

MLF

After trying to figure this one out I got some instruction by some guys over at Dell enterprise tech support (guys actually from the USA...shocking yet true) that they had found a hotfix that might address the problem I was having (since it had helped others with a similar problem).

I have applied it and so far it seems to have increased performance on our network.

They gave me this link http://support.microsoft.com/kb/924325/en-us

The title of the MS hotfix is...
"Network applications that use the NetBT keep-alive transmissions may not work correctly after you install the Windows Server 2003 Scalable Networking Pack on a Windows Server 2003-based computer"

Anyone know why this helps the laggg on the network from that file server? Could I apply it to other application and email servers and get more performance from them as well?

strange it only affected one machine...

edit> never mind..reread the thread :(
Thought it was just one machine...appears to be one app...


MLF

As Morgana~ says it seems to be one MS systems app.

So, if the server is running MS's Scalable Networking Pack it might help. Also check the clients to see if they have hotfix 916508 applied to them. If they do then they will also need the new fix applying to them.

As to why, well this is the fix that fixes the fix that fixed the fix that fixed the fix before the last fix that broke :D................ or something like that?

Basically the possibility for different environments is so large that they cannot test all eventualities...............I have gotten several of these fixes from MS over the years.

If your machines do not meet the criteria, don't apply the fix.............it may do more harm than good. That is why MS don't put them out on general release ;)

Ok so this problem has not gone away. It seemed for a moment that the computers were all doing fine, but then I noticed the laggg from the fileserver again.

It seems that any machine that is a couple switches away from where the fileserver connects or even just one switch over can become laggy when trying to get into the share drive on the server.

I never experience it because I am connected on the same switch, but It really is annoying when you have to wait for it. I found some critical NIC driver updates, but I have to come in early so I can install this without cutting people off from their files.

You shouldnt have any switches chained off other switches. Every switch should be home run back to your core switch. How many users do you have on the network?

Net2Infinity brings up an excellent point..

How many switches do you have...and how are they connected to eachother??

MLF

We have a total of about 60 users on our network. We have 10 Switches total and I really need to reorganize all of them so they all feed into one switch. Some of them are chained together which I am sure isn't doing any good.

In our second wiring closet/network rack we have 5 switches, but only 3 lines running from that room to the main server room. So what I have done is chain the two phone switches together and send that over one of the lines and the other two are shared between the other 3 switches (one of the three is chained into the other) that run all of the computes, consoles, etc.

In our main server room an unmanaged switch is where the internet, majority of servers plug in and then the other 4 switches attach to that (the two phone switches in this room are also chained together).

I would guess I should have one line run for each switch in our secondary wiring closet to our server room and to put a manged switch as the main switch and run each of the switches (including the phone switches) directly into this managed switch. Basically eliminate all daisy chains.

Well I see a problem...

AFAIK..there is the 4 hop rule... no more then 4 hops from switch to switch..

All wiring should be punched down to a central hub area...then hooked into a switch..with a max of 4 switches

switches can be plugged into switches to over come distance or other issues...but only up to 4 hops

This is definately an issue

MLF