bluewhaleCA
MIS
I've not had to dig this far before, and certainly could use help.
Our main office has about 50 PC/devices on it. 3-4 weeks ago we went with a new carrier, moving from 2 T1's to 3 T1's (bonded). We ran into some problems and twice the provider has had to call ATT out to resolve them.
We still have problems, but not ones that can easily be tracked down. We have a spike around the top of each hour, likely for our AV/Malware servers downloading updates. We checked and can not throttle the bandwidth in that program.
However we have had this spike going on for years: the prior carrier did not have a problem with it.
The problem manifests in the interruption of music streams for 5-10 minutes at the top of the hour. It interferes with Metaframe connections: sometimes causing a wait of 3-5 minutes for your desktop to respond again. The results of ping tests are ( done at same time to allow comparison ) 1 or 2 lost packets on the LAN (0%), and from 1 to 5% when pinging external sites for 5 minutes.
I've telnetted into our firewall and lost packets only one time in a dozen when pinging the carriers LAN and WAN ports.
I'm currently verifying if all users in this office are having problems or only a few.
I would have thought that if I consistently drop more packets ( much more ) when pinging external sites than on the LAN that the problem would have been in the Cisco router the carrier provided. ? If so what could I do to prove a problem exists? The best solution, taking the LAN off the Cisco and testing with a notebook is unfeasible. Perhaps early on the weekend, but not while people are working. But the problem seems to occur primarily when there is a load on the routers and you would not see that on the weekends.
Equipment: The carrier provided a Cisco router. We are using an HP 2900 and an HP 2824. I'm seeing intermittent excessive late collisions or excessive CRC/Alingment errors on half a dozen ports on the 2824, all but one going to printers or hall cams... And on the 2900 the only port that has errors is port1, which goes to the firewall.
Any thoughts on how to track down the cause(s) and 'prove' them will be appreciated.
Paul
Our main office has about 50 PC/devices on it. 3-4 weeks ago we went with a new carrier, moving from 2 T1's to 3 T1's (bonded). We ran into some problems and twice the provider has had to call ATT out to resolve them.
We still have problems, but not ones that can easily be tracked down. We have a spike around the top of each hour, likely for our AV/Malware servers downloading updates. We checked and can not throttle the bandwidth in that program.
However we have had this spike going on for years: the prior carrier did not have a problem with it.
The problem manifests in the interruption of music streams for 5-10 minutes at the top of the hour. It interferes with Metaframe connections: sometimes causing a wait of 3-5 minutes for your desktop to respond again. The results of ping tests are ( done at same time to allow comparison ) 1 or 2 lost packets on the LAN (0%), and from 1 to 5% when pinging external sites for 5 minutes.
I've telnetted into our firewall and lost packets only one time in a dozen when pinging the carriers LAN and WAN ports.
I'm currently verifying if all users in this office are having problems or only a few.
I would have thought that if I consistently drop more packets ( much more ) when pinging external sites than on the LAN that the problem would have been in the Cisco router the carrier provided. ? If so what could I do to prove a problem exists? The best solution, taking the LAN off the Cisco and testing with a notebook is unfeasible. Perhaps early on the weekend, but not while people are working. But the problem seems to occur primarily when there is a load on the routers and you would not see that on the weekends.
Equipment: The carrier provided a Cisco router. We are using an HP 2900 and an HP 2824. I'm seeing intermittent excessive late collisions or excessive CRC/Alingment errors on half a dozen ports on the 2824, all but one going to printers or hall cams... And on the 2900 the only port that has errors is port1, which goes to the firewall.
Any thoughts on how to track down the cause(s) and 'prove' them will be appreciated.
Paul
