jfo@mimsy.umd.edu (John F. Ormonde) (05/25/90)
Up until Monday we were cruising along with no problems on our 3/280,
16MB, 2-892 disks on a 451 controller serving 13 3/60s all running 4.0.3
Tuesday SUN installed 16MB, a second ethernet controller, another 892
disk and a 7053 controller. One of the existing disks was moved onto the
7053. I then added more variables to the equation by installing 4.1.
Since then things have been most unpleasant. A list of problems/symptoms
in no particular order follows:
1) Diskless client booting problems:
Boot: vmunix
.Size: 873240.......X..X.....{more '.'s and a few 'X's}
...+173280+218672 bytes
Statistics:
rpc: {various rpc stats}
le: {various ethernet packet stats}
continues with normal boot. This happens intermittently on random
clients.
2) Mailtool: aborts with the following error msg:
Mailtool not able to communicate with Mail!
(Perhaps Mail is not an executable appropriate for this machine.)
and mail takes a very long time to retrieve a 14000 byte mail file when
run on either clients and the server.
3) /usr/etc/etherd and traffic:
etherd dies due to rpc timeouts/failures
4) Clients get random NFS and RPC timeouts.
5) A find thru many (100s) subdirectories brought the system to its knees
for >15 min. It never took that long on the 451 controller and I thought
the 7053 was supposed to be faster.
I have tried my kernel, the generic kernel, pulling the 2nd ethernet
board, run sysdiag and calling our local SUN SW person and the SUN 800
number. I am seriously considering going back to 4.0.3 unless I can get
some help from somewhere.
Please reply via e-mail.
Thanks
john ormonde
jfo@mimsy.umd.edu