> Is there a way to get info on how busy a Portmaster 2e is?
No - even with all 30 ports going FULL speed it can keep up.
> Here's why I ask. I've been trying, and I mean trying, to play
> Quake via my portmasters for a while. I can connect and play and such,
> but I have been experiencing these freezes in the game periodically.
> Usually, at the worst possible times and usually when the server is
> trying to send a lot of data. It seems as if packets are piling up on
> the Portmaster. When the game unfreezes, everything moves at a much
> accelerated rate.
>
> I also run a bank of Shiva LanRover/E Pluses. When I played using
> the Shivas, it was almost as if I was playing via a LAN. A little
> slower, but MUCH better than playing via the Portmasters.
These are using USR Courier V.Everything modems as you mention below
right?
> I know that this is hardly scientific, but it is indicative of a
> problem. So, in the interest of providing better service to my
> users(And improve my game(-; ), I'm trying to figure out what's going
> on.
Ok.
> Both dialups are connected to our LAN the same. Both are using USR
> V.everything modems. Both are connected via Cisco routers.
Okay - how are the modems set up (init string and dip switch wise)?
How are the ports on the Portmaster and then on the Shiva setup? Flow
control settings? Speed? Etc.. What is the client modem(s) that you
dial in with and have problems on the PM2's with?
> The LAN segment the Portmasters are on is not overloaded. I put a
> LAN analyser on it and usage peaked at 15%. Average usage is 5%. Even
> less now that the students have gone home for Thanksgiving.
Okay. How do the collision rates look? (type show netstat and on the top
it will say something like this):
Name Ipkts Ierrs Opkts Oerrs Collis Resets Queue
ether0 7211051 0 7265449 0 15935 0 0
What about the error counters on the ether0?
> I tried cruising through the SNMP dumps, but didn't find anything I
> could use. Is there anything there that I could use?
For what? Nothing for CPU usage..
> What about the "show mem" command? What does the series of colon-
> seperated numbers mean?
Memory fragments, example:
System memory 1048576 bytes - 981924 used, 66652 available
2048:6 4240:3 1152:1 128:10 640:8 144:7 80:9 96:6 64:0 176:8 16:110 160:8
32:61 48:8
System nbufs 1400 - 76 used, 1324 available
Which means there are 6 2048 byte fragmets, 3 4240k fragments, 1 1152k
fragment, 10 128k fragments, etc.
How do the nbufs look when it is slowing down? What about the bytes
pending on the port (show sXX or show all) when it slows down?
- Steve
- Systems Manager
- Community Internet Access, Inc.
- Gallup and Grants, New Mexico
-
To unsubscribe, email 'majordomo@livingston.com' with
'unsubscribe portmaster-users' in the body of the message.