OK, a quick update. I tried building another machine with the same hardware
(minus the realtek) and it works just fine. Its not nearly as busy as the
one that failed. However, I did notice that the quota.users files on the
two partitions had strange time stamps of the last access. Is it possible
that the OS was unable to write to these corrupted files, and this is where
everything was holding up ? Anyways, I blew away those files and will try
a reboot in the early hours to see if that makes a difference-- i.e. with a
freshly created quota.user file.
---Mike
At 08:45 AM 1/2/02 -0500, Mike Tancsa wrote:
>OK, I know I will get slapped for providing such little information (to
>start), but this morning I had a bit of fun with upgrading a Oct19 kernel
>to a Dec31st kernel. Getting rid of
>
>#options DUMMYNET
>#options IPDIVERT #divert sockets
># Disk quotas are supported when this option is enabled.
>#options QUOTA #enable disk quotas
>
>allowed me to have the machine function in a normal state. The problem
>state was that I was unable to login to the system in multi user mode.
>There *seemed* to be a lot of processes blocking on something. I dont
>know what as I was unable to login. I thought at first it was that syslog
>problem with /dev/console, but I rebooted with syslogd off and I was able
>to get to a login prompt briefly. After that the system became unstable
>with too many files open I had to reboot.
>
>I am posting this in case anyone else has seen something similar.
>
>dmesg below. Kernel config is basically GENERIC with uneeded device
>drivers removed. MAXUSERS is set at 256.
>
>As the box is production, I cant fiddle with it right now, but I can come
>back in during off hours and do whatever would be useful to try and track
>this down. I do have another box with quotas enabled, but its nowhere
>nearly as busy, particularly as I tried to bring up quota support on my
>mail partition which is fairly busy. (mount with soft updates
>enabled). Anyone see anything similar ?
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-stable" in the body of the message
