It was requested by several people that I post a summary, so...
Thanks to the many people who replied. Here is the original question:
------------
Date: Mon, 03 Dec 90 18:02:36 EST
Subject: Logins in Disk-Wait
Help! Our machine keeps crashing...
As we approach the end of the semester, the load on the network has been incr
- easing. It seems that when the network traffic is at its highest, processe
- s on the file server (mostly logins) go into disk-waits states.
These of course cannot be killed, and simply add to the load that is there al
- ready. This causes the load to spiral until the server dies. It crashed o
- vernight last week with a load of over 35!
Does anyone have any suggestions as to what we can to control this? Is there
- any way to kill disk-waiting logins?
Our setup:
Main server: Sparc 4/470.
YP Slave: Sparc 4/280.
around 100 workstations, mixture of Sparc 1s, Sparc 1+s, and 4/110s. SunOS 4
- .0.3 on about 50%, 4.1 on the other 50%.
------------
Well, the responses were many and quite helpful. There were a few who said that we need more servers/gateways/etc. to help relieve the load. Well, yeah that's true. We're working on it.
Some said that the nfsd bug patch would help. Well, we've already installed that so that isn't the same problem.
By far the majority of the responses were about the quotas. It seems that automounter defaults to "quotaon". Ahhh. I've changed all of the automounter maps on our machines, and the problem has not reoccurred. Yet. We shall see.
Thanks again to all those who responded!
- Bo
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:00 CDT