[913] in NetBSD-Development
Re: Status update
daemon@ATHENA.MIT.EDU (Charles Hannum)
Thu Jul 27 12:17:23 1995
Date: Thu, 27 Jul 1995 12:10:00 -0400
From: Charles Hannum <Charles-Hannum@deshaw.com>
To: ghudson@MIT.EDU
Cc: pc-dialup@MIT.EDU
In-Reply-To: <9507262308.AA15278@pan-galactic-gargle-blaster.MIT.EDU> (ghudson@MIT.EDU)
We ran into several possible hardware problems with the machine:
* "le0: crc mismatch" errors
This is not a `problem' with jaws, per se. It's a problem with your
network. You're getting corrupted packets. If the network is to
spec, this should never happen.
* Occasional "excessive collision" errors from le0
This is another network problem. This particular error causes an
outgoing packet to be dropped, and the higher level protocol has to
retransmit it.
* AFS performance seems on the slow side compared to other
NetBSD machines; this may be my imagination (I noticed it
while copying over the srvd, which isn't the kind of
operation I do very often).
You were creating lots of little files, which is slow.
* Very occasional timeout warnings from bt0 on the internal
disk; these don't cause problems
Not sure about this.
* After any significant amount of activity on the external
disk, the machine hangs with several iterations of:
Jul 26 15:15:43 jaws /netbsd: bt0: mbi at 0xf8745dc8 should be found, stat=00..resync
I doubt that is the actual problem. The message is merely
informational; it has to do with a race condition that is expected.
There are two actual problems I know of:
1) I saw a timeout on sd0 once. This shouldn't cause any serious
problems.
2) When there is both local file system and AFS activity, the machine
occasionally hangs. This needs to be investigated.
