Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 27 Jul 1995 11:40:44 -0600 (MDT)
From:      Brad Midgley <brad@pht.com>
To:        hackers@freebsd.org
Subject:   2.0.5-950622-SNAP on a big machine
Message-ID:  <Pine.LNX.3.91.950727113323.31701C-100000@exodus.pht.com>

next in thread | raw e-mail | index | archive | help
Hi all,

I just upgraded from a 2.0 snap to this version on our ftp/www site.  I 
was mainly upgrading for stability and so we could have a perl that wouldn't 
fail its self-tests.

machine: pci p60 w/96mb ram, 5 4gig drives (quantum, seagate) and 2 2gig
drives (all scsi), buslogic scsi, 3com ethernet. kernel configs:  users=128,
options "NMBCLUSTERS=2048", disabled a lot of things, including
bounce-buffers, ide, cdrom, iso9660. 

some problems now:

 after only about 12 hours of uptime, the system virtually locked up with
messages about bt0: buffer full (the system was still pingable, great). 
None of the errors were logged (I should have expected that) so I don't
have them verbatim. 

 it appears that freebsd only sees 64 of the 96 mb ram on the machine.  the
bios self-tests on boot see it all.  How do I monitor ram/swap usage on a
running system? 

 the machine had an smc card which wasn't recognized by the new system.
to be more accurate, I prepped the boot drive on a machine with a 3com 
and then changed /etc/sysconfig's 

 network_interfaces="ep0 lo0"
   to
 network_interfaces="ed0 lo0"

that didn't work in the server itself with the smc, so I swapped the cards
themselves around and switched it back.  3com support looks much better btw. 

Sorry if these are things which have come up--I had to drop this list 
because my mailbox was buried with new mail all the time.  Please cc 
replies to me, because I may miss them (just resubscribed to the digest 
version today)

Brad



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.LNX.3.91.950727113323.31701C-100000>