Applied losing network connection

Lynne Desrochers · March 07, 2014, 11:39:14 AM

I recently had two people start working in my office that are working with an online sharing program, syncedtool from Anchor. They wanted to scan at their desktops directly to the syncedtool web site. We are using an outside tech to manage this. They are only here a few days a week. It's just getting started this week and they haven't really scanned much.
This week, i've been having people get stuck in applied and getting the "server not available" error in various areas of applied. It wasn't everyone, so I haven't been able to pinpoint it. Today around 10:00 am 13 people got those same error messages again. When I asked if anything was scanned today it was right around 10:00 when everyone got the errors.
I want to blame the scanning directly to the website. I know that in our remote office we have them scan to the local desktop, then copy it over to the network. We do not have them scan directly to the network drive.
Anyone have any thoughts, suggestions or input?

Mark · March 07, 2014, 11:45:11 AM

Think I'd need a little more info about the environment and the workflow process to troubleshoot this.

Are these remote users that are scanning at their desk, t hen transferring over the network to a server before uploading?

What is this syncedtool? Is there a VPN involved or anything like that?

Lynne Desrochers · March 07, 2014, 11:51:18 AM

The ones using the synced tool are working here on my network. They are scanning from their desktops directly to the syncedtool website, so we mapped a drive to the syncedtool website and have the scanner pointing there.
No VPN, it seems to me like DropBox. just a bunch of files that you can access from anywhere.

Mark · March 07, 2014, 11:52:49 AM

Ok, got it. I did Google it too and I agree it is like a dropbox type thing.

So, who is losing network connection when this happens? People in the same office as the syncedtool users, remote users..?

Lynne Desrochers · March 07, 2014, 12:03:04 PM

Other people in the office. The people using the syncedtool are not using Applied. I just found two errors from yesterday and I was able to find scans around the same times.

Mark · March 07, 2014, 12:04:54 PM

Ok, so it's people in the same physical location and this is the same physical location of the TAM server?

And also, are these all virtual desktops we are talking about of physical local machines?

Lynne Desrochers · March 07, 2014, 12:13:36 PM

all people in the same location and all have physical local machines.

Mark · March 07, 2014, 12:19:49 PM

I wonder if this might be related to the amount of traffic on a switch or switchport then. Can you make any kind of relation as to where the machines and server are connected to a switch?

Maybe there is a bottleneck. Do you have multiple switches? Is the TAM server connected to one switch and BOTH the sync users AND the TAM stations losing network connection connected to a switch OTHER than the switch the TAM server is connected to?

Also, what is/are the switch(es) in play here?

Lastly, is the error they are seeing coming from TAM?

Lynne Desrochers · March 07, 2014, 12:33:19 PM

The error is from Tam, various areas depending on where they are at the time. ex: An error occurred while displaying list., then under desc: Could not read at least 8 bytes at offset 0 in "F:\TAM\CLIENT.NDX": Error 64. the specified network name is no longer available.
and other various ones that end with network name is no longer available.

Jeff Zylstra · March 07, 2014, 12:51:44 PM

If a network connection is being lost, I would hope that an error message would be logged somewhere, like event viewer. Have you looked at the System log in event viewer for both the machines that lose the connection, and also the machines that are doing the scanning? I'm also wondering if there aren't errors in the Event Viewer - Applications for both the machines that are scanning, and also the machines losing a connection.

I don't know if there is a third party utility that could analyze the traffic patterns and log errors or conflicts, but it would seem to be useful in this situation. Don't know if Wireshark or something similar would help or not.

Mark · March 07, 2014, 12:55:42 PM

If only TAM is reporting the error, I wouldn't suspect the full network connection is going down, i.e. the balloon isn't popping up saying that a "network cable is unplugged".

I would almost suggest it could be a DNS issue, but I'm leaning more towards the switch at this point.

See what kind of relationship you can determine with the switch as I mentioned in my previous post. Maybe you need to create a bigger trunk between two switches, or just move some things around to even out pressure on one port. Something to that effect. QoS even prioritizing traffic to the IP address of the TAM server or limiting bandwidth to the sync program.

Lynne Desrochers · March 07, 2014, 03:03:14 PM

Only Applied is showing the error, but their email is on their local C drive and if the website they were viewing skipped a beat or two they might not notice. There aren't many other things that are specifically on the server that need the constant connection like applied. Am I off base in that thinking?

Mark · March 10, 2014, 10:11:06 AM

Is Citrix in the picture at all?

Lynne Desrochers · March 10, 2014, 10:34:53 AM

I do have two people on citrix and they got an applied error, but it also could have been that the citrix is on the same switch.
I think I'm going to test out the scanning tonight at 5 and see if I can re-create the issue.

Jeff Golas · March 10, 2014, 10:43:43 AM

I'd be looking at a few things:

Bandwidth - is it being chewed up? (I'd be willing to bet that dropbox-like system is really p2p thats STORING things on your systems as well as letting you store elsewhere)

Are the computers running this software the ones having issues? Could also be running out of network resources (again prob tons upon tons of connections getting opened, do a netstat -a).

Router - even if you have the bandwidth, if it suddenly spikes huge, the processor in whatever router/switch you're running may not be able to handle it.

Jeff Zylstra · March 10, 2014, 10:47:19 AM

Quote from: Lynne Desrochers on March 10, 2014, 10:34:53 AM
I do have two people on citrix and they got an applied error, but it also could have been that the citrix is on the same switch.
I think I'm going to test out the scanning tonight at 5 and see if I can re-create the issue.

We had a remote employee who used to do a lot of scanning from her house in Florida, to our office in Michigan. She scanned to a mapped drive that was facilitated via a VPN connection. She was a heavy scanner, but we never had an connectivity issues that caused TAM or other network issues.

Along with checking network and switch issues, I would also check and see if any other program or protocol is running in conjunction with the scanning operations. You might be surprised at "helper" programs are running in addition to a straight ftp or http connection.

Mark · March 10, 2014, 10:51:17 AM

No router in play here since it's local LAN traffic on the same subnet.

I am leaning towards the switch. If you can limit the bandwidth the sync app uses or give the TAM server priority (which would require QoS) that may help.

Try putting the TAM server on a different switch and see if that makes a difference. 12 users doesn't create a lot of traffic though. It's gotta be the sync tool saturating the line somewhere.

How many switches do you have?
How many ports are used to link the switches?
What kind of switches?

Lynne Desrochers · March 10, 2014, 05:52:11 PM

Jeff G- I don't think it's bandwidth, since 50 out of 52 are local, including the girl who was trying to scan. Computers running this software are not running applied, yet. I feel like it's the switch.
Jeff Z - interesting that she scanned to a mapped drive that was facilitated by a VPN, I think our initial issue with the remote user scanning to a mapped drive was QoS was not turned on properly and it ate up the phone bandwidth.
Mark - I think it is the switch as well. We have 3 HP switches, two 48 port and one 24 port. When i did the initial check, not everyone effected was on the same switch, but those things are hard to follow, so I may have been wrong. It's not just 12 uses, we have 54 people, only 12 complained that day about having an error in applied.

Would i be crazy to give everyone a heads up and then tomorrow at say 4:30 to run a few scans with the original configuration and see if I can replicate my issue? One last clue, I have an incremental backup running through out the day, and this has been running for months. I'm wondering out loud if it could be the timing of it all.....

Mark · March 11, 2014, 10:51:55 AM

It very well could be the timing. How does the backup run during hours? Is it backing things up over the network? It's not backing up TAM while users are in, is it?

I just saw "12" in your signature and thought it was number of users

Are these switches gigabit? Fairly new? They are not old HP "hubs" are they?

Edit: Re-reading your post, you mention "phones" so I'm guessing these are NOT hubs, lol