Why would resetting all switches in our office network solve the problem?

By sirjeff2000 ·
I have a small office network. Because of location of the offices, we connect workstations through several switches. We have a windows 2003 server. We are not using domains or active directory. The network is primarily used to access the internet. We use a Fortinet Content filter/firewall. DHCP is provided by the Fortinet device. The server is mainly used for storage and a centralized accounting database.

I have one user setup so that her documents folder is actually on the server and her workstation has a mapped drive to that folder. The server is setup to mirror it's two drives for redundancy.

Occasionally, that user will experience a situation where she will open a file or create a new file most commonly with MS Publisher and when she attempts to save the file will get an error message that she is out of disk space or the dreaded hourglass and cannot save to her documents folder on the server. I can go to the mapped drive and see the folders there but cannot save the file.

When this happens it seems to cause a slow down of all traffic on the network. The only way I have found to correct the problem is to turn off all switches and internet connections and then power them back on. Once all connections are restored, the file can be saved. It's not necessary to reboot any of the workstations or the server.

It's as if something goes into some sort of loop.

The switches are unmanaged and are a combination of 3com and Netgear devices.

Any help in this situation would be greatly appreciated.

This conversation is currently closed to new comments.

Thread display: Collapse - | Expand +

All Answers

Collapse -

Most likely either a device is 'jabbering' or is faulty

by robo_dev In reply to Why would resetting all s ...

If a PC is misconfigured, such as if the Ethernet port is configured for half-duplex, or there is an Ethernet port that is bad, or a bad cable, or even software misconfiguration or malware, then resetting the hardware can do several things:

a) it clears the forwarding table memory of the switch, which maps what switch-port goes to which mac address on which port. Sometimes if there are errors caused by a user plugging a hub into a switch, or even by wireless devices, the forwarding table can get hosed.

b) it may clear up a temporary condition, like a virus on a PC that's flooding the network with connection requests. By disrupting the process, it fails, and it takes it time to recover.

c) A faulty device may be overheating, or even there may be a 'memory leak', such that a router/firewall process is allocating memory, then not doing garbage collection....then resetting it fixes that problem.

To fix your problem, connect a hub between the user's PC and her switch port, then install the free WireShark sniffer utility to monitor the traffic. This should tell you what LAN problem is occurring fairly easily.

Collapse -

Reponse To Answer

by sirjeff2000 In reply to Most likely either a devi ...

The idea of "jabber" was something I had considered. Using WireShark is a possibility but because of how intermittent it is doesn't lend itself to easily to finding the culprit. I'm afraid, I would have to dedicate a computer to run the capture program. I know I can filter the capture packets but haven't a clue what to look for. I understand the capture packet process and can read rudimentarily but an expert at captured packets I am not. I would appreciate any more detailed information you can give. I'm in the process of mapping out the network to see if there is any way to isolate sections and see if it makes a difference.


Collapse -

we have a Mac server

by Sue T In reply to Why would resetting all s ...

that has those problems also. We found that we only have to reboot that server and we are good to go until the next time it decides to hiccup.

Related Discussions

Related Forums