Jumbo frames will have little benifit to you if all your end devices are not set to use jumbo frames. Is it possible that with only a single gigE link between your switches that the clients are saturating it. Any idea what sort of aggregate speeds the clients are pulling when this happens?
Well if you have an extra port on both switches you could setup a LAG between them. However you may not see the full benifit of the lag with so few mac address pairs on the switches. But it might be worth a shot.
It is possible that we could be saturating the line because our user download TONS of data from our SQL machine. We are an auditing firm and our users actually might need to see 100,000 records at a time for checking duplication within the data over time. All of the clients are at 1 gig so they could be filling the pipe entirely.
I also have a few 100 megabit clients hooked up to the switch where the servers are hooked and everything on both side is set to Auto mode. I have also turned off the gigabit advertisement on these ports to no avail....
I have replaced the 2716 with an old switch that we used to use to narrow down the problem and see if it's really the 2716 or not.
I don't think that the switch is actually crashing per-se, but my users end up with a "Disk or network error" within Microsoft Access. This leads me to believe that their connection to the file server has been severed. Previous to the connection dying, the application performance gets worse until it's to the point of unusability.
Maybe i didnt understand your original problem then, is the switch actually crashing? I though you were just having application performance issues, which could easily be from a saturated link.
I think I found the source of the entire problem....somebody replaced one of my Cat5e ethernet cables with an old CAT4 cable. I don't know why it was replaced or who did it, but after replacing it and putting the 2716 back into service, we haven't had a problem yet. I'm hoping this cures the whole situation!
DRNO10
184 Posts
0
August 21st, 2008 13:00
Jumbo frames will have little benifit to you if all your end devices are not set to use jumbo frames. Is it possible that with only a single gigE link between your switches that the clients are saturating it. Any idea what sort of aggregate speeds the clients are pulling when this happens?
jsacrey
6 Posts
0
August 21st, 2008 14:00
Gotcha, I do have extra ports and will try that when I get a chance.
Do you think the switch port is being overloaded and crashing the switch itself?
DRNO10
184 Posts
0
August 21st, 2008 14:00
Well if you have an extra port on both switches you could setup a LAG between them. However you may not see the full benifit of the lag with so few mac address pairs on the switches. But it might be worth a shot.
jsacrey
6 Posts
0
August 21st, 2008 14:00
It is possible that we could be saturating the line because our user download TONS of data from our SQL machine. We are an auditing firm and our users actually might need to see 100,000 records at a time for checking duplication within the data over time. All of the clients are at 1 gig so they could be filling the pipe entirely.
I also have a few 100 megabit clients hooked up to the switch where the servers are hooked and everything on both side is set to Auto mode. I have also turned off the gigabit advertisement on these ports to no avail....
I have replaced the 2716 with an old switch that we used to use to narrow down the problem and see if it's really the 2716 or not.
Any other ideas?
jsacrey
6 Posts
0
August 21st, 2008 15:00
I don't think that the switch is actually crashing per-se, but my users end up with a "Disk or network error" within Microsoft Access. This leads me to believe that their connection to the file server has been severed. Previous to the connection dying, the application performance gets worse until it's to the point of unusability.
DRNO10
184 Posts
0
August 21st, 2008 15:00
Maybe i didnt understand your original problem then, is the switch actually crashing? I though you were just having application performance issues, which could easily be from a saturated link.
DRNO10
184 Posts
0
August 27th, 2008 17:00
jsacrey
6 Posts
0
August 27th, 2008 17:00
Hey DRNO10,
I think I found the source of the entire problem....somebody replaced one of my Cat5e ethernet cables with an old CAT4 cable. I don't know why it was replaced or who did it, but after replacing it and putting the 2716 back into service, we haven't had a problem yet. I'm hoping this cures the whole situation!
Thanks for the advice in the meantime,
Jim
jsacrey
6 Posts
0
August 27th, 2008 18:00
No kidding! Thanks again!