From: Tom Grassia (tgrassia@sfnewmexican.com)
Date: Thu Jan 20 2005 - 14:20:12 EST
Hello,
We have a very ghetto Sun Ultra 2 (creator) being used as a critical
database server. Don't ask me why that is, I just got here. I've lost
count of the number of external SCSI hard drives connected to this
thing. As near as I can tell, there are five, the wires of which are
entangles with external drives for a second Ultra 2 being used as a second
critical db server.
I've started to notice SCSI errors, which are leading me to think some of
the drives might be going bad.
To kill two birds with one stone, I bought a refurbished StorEdge D1000. I
figured, "I'll attach this JBOD, get it configured with Disk Suite like the
other hard drives here, then remove the external drives from the
server. That way I'll have fresh working drives all stored in one nice
box, and we'll go from 'ghetto' to 'working poor'."
The D1000 arrived. I brought the system to single user, executed power off
from the prom, then powered off the external hard drives. After that, I
cracked the case and installed the differential SCSI card for the D1000. I
put the cover back on, reattached the drives, then attached the D1000 to
the U2 and powered back up.
The server wigged out on me. The screen flooded with all sorts of errors,
got up to what it thought was multi-user but wouldn't allow anyone to log
in. I believe this is because the home directories didn't mount
properly. The system prompted for a login, but hung after a username and
password were entered. Then the login disappeared, so I guess it couldn't
find a home directory.
I unattached the D1000, then restarted. The system hung again. I powered
down, then switched around a few SCSI connections. I figure I probably had
a couple drives attached to the wrong plugs in the back of the server. The
server came back up into multi-user mode and I let folks get back to
work. We have a few small windows for downtime.
I'm preparing to do this again, and I really want to make sure I can bring
up the server with the D1000. I've got the syslog for the time period when
we attempted to do this.I don't want to attach it, because with all the
errors that tripped off and repeated it's about 220k, and I don't want to
spam the list like that. Neither do I want to try and trim the log and
possibly delete the error message that would give the clearest picture of
what's going wrong.
Would just having the D1000 hooked up when the other drives were in the
wrong spots cause a ton of error messages? When I changed them around with
the D1000, I didn't get anywhere near as many errors. If anyone would like
to see the log, I'd be more than happy to send it out as well.
Thanks for your time,
--Tom
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:30:02 EDT