luminousmonkey.org

OpenSolaris and no menu screen boot issue

Friday 17 October 2008, 13:44 WST —

I like OpenSolaris, ZFS is very cool, and I’ve toyed with zones a little, and I’m yet to really try DTrace.  However, when it comes to ZFS, there’s one thing that’s a bit of a worry, the complete and utter loss of your data.  Now in theory if you’ve got a redundant pool, then your dataloss should be very minimal, however, certain things can happen that you should be careful playing with things that are cutting edge.

What the hell am I on about?

Well, on my home server, all by itself, it decided that it would no longer boot.  GRUB would load, however the menu wouldn’t, no spashimage, no menu selection, just the grub command line.  I really couldn’t be bothered to try and fix it to be honest, and it gave me a chance to document how to get MediaTomb, etc, setup for it.

The problem, is pretty much described in this OpenSolaris forum post.

Anyway, today, I had the same problem with our email server at work, I rebooted, and the machine wouldn’t boot.  Because it wasn’t easier to start from stratch, I had to try and nut the problem out.  Unfortunately I don’t even know what the problem is, I’m just not that good, I had already tried directly entering grub commands to get it to boot (getting the same result as in the forum post).

Luckily, using Google, I found some pointers to get myself sorted.  Actually, this post, they even made a wiki entry for it here.  First time I tried it, it didn’t solve the problem, however, I thought I would also try it on the older boot enviroment… success!!! Sort of….

Initially, I had only done the update on rpool/ROOT/opensolaris-snv_95-1 (it is updated to snv98 despite the name), however, after also doing the procedure on rpool/ROOT/opensolaris-snv_95, I got my grub menu back!! Woohoo! Oh, wait… er… kernel panic….

So after much wailing, thinking, and googling, I really had nothing else to try, but I had noticed that there was a opensolaris-work-2 boot environment I had forgot.  So basically I followed the procedure for all three of my boot environments, I also just left the zpool imported (I was using a snv98 LiveCD) and rebooted the machine.

Success! So now the server is up and running again, but it’s things like this that really scare me about using OpenSolaris and ZFS.

Update:
I’ve recently found out that the RAM in my home server is faulty, no matter how good a filesystem is, it still can’t stop the hardware from screwing it up.  However, it has to be said that, despite the flakey hardware I’ve only lost one, or two files.