?

Log in

No account? Create an account
23:24 < haroot> isaacs: Ru running gentoo in ucsc? 23:24… - CERisE's Testing for L

> Recent Entries
> Archive
> Friends
> Profile

March 14th, 2004


Previous Entry Share Next Entry
02:34 am

23:24 < haroot> isaacs: Ru running gentoo in ucsc?
23:24 < isaacs> haroot, no. i'm in scotts valley
23:25 < haroot> heh
23:25 < haroot> If I only knew where that was exactly
23:25 < haroot> how do you like gentoo so far?
23:26 < isaacs> great. love it. can't believe cerise_ had to break my legs to 
                get me to try it
23:26 < haroot> lol


In other news, Arrakis (relsquis router) lost the last sector on its hard drivve today. That doesn't seem so bad...except for the part where it's using XFS and XFS won't mount if it's log is unreadable. As a result, that computer is more or less dead.

Fortunately, I was able to boot from the Gentoo CD and setup all the NAT stuff through iptables. I even compiled udhcp on l, unpacked it on Arrakis, and started it up again.

I have a 4 gig drive at home that I'll bring up next time and slap in. Hopefully all that compiling won't all be for naught. I imagine I'll be able to simply cat the partition over and handle things on the larger drive.

At least I won't have to be quite so stingy. I was clearing out /usr/portage/distfiles, /var/tmp/portage, and /usr/share/doc every time merger completed. GNOME and KDE could probably find room on that system to impress the mundanes.

Fortunately appear (the portage frontend I wrote) makes it pretty easy to drop in and forget about.

(3 comments | Leave a comment)

Comments:


[User Picture]
From:abadman
Date:March 14th, 2004 12:07 pm (UTC)
(Link)
I imagine I'll be able to simply cat the partition over and handle things on the larger drive.

Just rsync it. I've done that a number of times even when making major hardware changes and/or filesystem changes. Just setup your filesystems mount them in your prefferred arranegment then do something like

/usr/bin/rsync -va --exclude /proc --exclude /mnt/new-drive / /mnt/new-drive/

Then just modify fstab, mkdir /mnt/new-drive/proc install a bootloader or modify the existing one if your keeping that drive and you in like flint.

Of course in situation where your swapping a lot of hardware you sometimes have to mess with modules and XF86config etc., and of course you have to make sure the compile flags on the old system match the processor of the new, but it's basically the same thing.

I stumbled across this journal by way of
[Error: Irreparable invalid markup ('<lj-user="gentoo">') in entry. Owner must fix manually. Raw contents below.]

<i>I imagine I'll be able to simply cat the partition over and handle things on the larger drive.</i>

Just rsync it. I've done that a number of times even when making major hardware changes and/or filesystem changes. Just setup your filesystems mount them in your prefferred arranegment then do something like

/usr/bin/rsync -va --exclude /proc --exclude /mnt/new-drive / /mnt/new-drive/

Then just modify fstab, mkdir /mnt/new-drive/proc install a bootloader or modify the existing one if your keeping that drive and you in like flint.

Of course in situation where your swapping a lot of hardware you sometimes have to mess with modules and XF86config etc., and of course you have to make sure the compile flags on the old system match the processor of the new, but it's basically the same thing.

I stumbled across this journal by way of <lj-user="gentoo"> and find your high score on the geek index intriguing. I find myself constantly avoiding making uber-geeky posts since I figure no one will read them. Its also very cool that you write code. I'm drifting in that direction myself with more of a sysadmin background. Anyway, I'm adding you.
[User Picture]
From:testing4l
Date:March 14th, 2004 03:37 pm (UTC)
(Link)
That would be a fine idea...except that the partition in question doesn't mount. The last sector on it is dead and that's where XFS keeps its log. It won't mount with an unreadable log and xfs_repair can't zero the log on a bad sector.

I think it may mount with -o norecovery. Failing that, I could probably define a different logdevice, use xfs_repair to wipe it, and move from there.

Failing that, I've always liked the ability to use cat to transfer raw data ; ) I believe XFS can be dynamically resized into a larger filesystem, so I plan to cat it over, zero the log, mount it normally, resize the partition, then resize the filesystem.

All the same, I'm not inclined to mess with it until the new drive gets here (we actually just bought one via ebay), lest it gets worse.

I actually went the opposite direction. I coded first, then one day, I got interested in this Linux thing everyone was yammering about. Having no resource other than one of the Walnut Creek Slackware CDs, I stumbled through the install and figured it out as I went along.

About 10 years later, a number of OSes later, and I'm running a free shell account server (l.armory.com).

In any case, nice meeting you ; )
(Deleted comment)
[User Picture]
From:testing4l
Date:March 15th, 2004 09:15 pm (UTC)
(Link)
It isn't. What I was referring to was stock portage with the line

"rm -rf /var/tmp/portage/* /usr/share/doc/* /usr/portage/distfiles/* >& /dev/null"

which was put right before the ending print line of merger.

You'll get credit and no, that addition isn't done, but I have more people asking for it, so I'm slightly more motivated : P

As for monkeys...Maybe when monkeys fly out of my ass!!!

Oh wait...

Hmmm.

> Go to Top
LiveJournal.com