Some of you remember that we have upgraded some of our servers to Fedora last week.
Since then, I have not been sleeping much. The server we upgraded is having these weird kernel panics, even though we tested it. The server reboots fine, and comes back up after five minutes, but it continues to happen every other day or so. So this weekend, the server folks and us did tons of tests over the middle of the night. We think we found the issue, and we are hoping this server holds up, so that we can upgrade the others using the same method.
But I have not been sleeping. Besides for being up with the server upgrades, I have this sick feeling all the time. The feeling of worry that the server may be down from a kernel panic. It is horrible, especially when you have to go offline for a 24 hour period.
Last night I was up since about 2am. I just hope all goes well with the server today, so maybe I can sleep well tonight, for the conference tomorrow.
All in all, horrible to move servers and homes in the same week.


Comments
DIDNT SLEEP
u are crazy
dont worry it will wokr out
and let ur brother do it and let him stay up
did he??
Posted by: server | October 15, 2007 10:45 AM
Really... I thought you had employees to worry about things like that. Get some sleep and make them worry about it. :)
We've got a server that does a panic about every 2 months due to the SCSI controller getting a bad request. The fix? We scheduled it to reboot weekly. Haven't had a problem since; knock on wood.
Posted by: Brian Mark | October 15, 2007 11:08 AM
Yea, we have people we pay a lot of money to worry about it.
But when it comes down to it, it is my business, my clients, I care and worry.
Posted by: Barry Schwartz | October 15, 2007 11:10 AM
I didn't mention it sooner because I thought your team had worked through it. I had a similar problem on a dell server using fedora a few years ago. The server is a dell poweredge with broadcom NIC's and we were using Fedora. I had two additional Intel NIC's and turned those on, turned off the broadcom's and the problem went away. Months after I posted in a discussion board for help, someone responded to me with the following solution:
=====
I fixed my crash problem disabling "dri"
capabilities to X session at xorg.conf.
Hope next version of kernel-smp and ATI 7000 will solved it.
=====
I stopped using Fedora a couple of years ago, in favor of Debian. I have not looked back since. Debian has things in different places than Fedora, but once you figure out where everything is it is so much easier to update, upgrade, and install things. When I was setting up a load balanced cluster, I got stuck after two days with Fedora but in Debian after a few hours I was done.
my 2 cents
Posted by: Merrick | October 15, 2007 2:12 PM