In practice, the backup takes an hour or two once every 6 monthsHi,
In message ,
I wrote
| Basically I keep a single fdisk partition containing the entire disk,
| but two sets of OpenBSD root, usr, and now var partitions inside that,
| both sharing /home and /data (where I keep my user files):
| wd0a root fstab mounts root, usr, var, home, data
| wd0b swap
| wd0c entire disk
| wd0d root2 fstab mounts root2, usr2, var2, home, data
| wd0e var
| wd0f var2
| wd0g usr
| wd0h usr2
| wd0j home
| wd0k data
|
| I use the standard OpenBSD bootloader; typing "boot wd0a:/bsd"
| (or just doing nothing and waiting for the 5 second default timeout)
| boots the wd0[aeg] set of partitions, while "boot wd0d:/bsd" boots
| the wd0[dfh] partitions.On Sun, 25 May 2008, Leo Baltus wrote:
> overwrite filesystems you normally use,
The backup (copy a,e,g to d,f,h) doesn't overwrite filesystems I
"normally use"; it overwrites a *backup* set of filesystems which
I typically haven't even mounted for some months. That is, my wd0a
/etc/fstab mounts *only* the a,e,g,j,k partitions, so when booting
from wd0a, the d,f,h partitions are not mounted. Similarly, my wd0d
/etc/fstab doesn't mount the a,e,g partitions. (The idea is that
having the "other" set of partitions unmounted keeps them safe from
a great many rm-in-the-wrong-directory type sysadmin blunders.)
> hoping the backup is a copy
This is a very important point. I completely agree, backups are
useless if I can't rely on them. So I test them as best I can
before proceeding with an upgrade/reinstall on the a,e,g partitions.
For example, for my 4.2-stable --> 4.3-release transition, my sequence
was:
1. copy 4.2-stable a,e,g --> d,f,h, run installboot on d
2. reboot from d,f,h and use laptop normally for 4 or 5 days to make
sure that the d,f,h 4.2-stable works normally
3. fresh install of 4.3-release on a,e,g
4. reboot from a,e,g 4.3-release, work through my usual post-install
checklist of config file edits, packages, one or two ports, etc etc
5. use laptop normally running a,e,g 4.3-release; track down & fix any
remaining glitches that come up (I might hand-mount d,f,h read-only
for a while, just to have them handy for glitch-fixing)
After step 2 I don't have to just "hope" the d,f,h backup is a copy I
can rely on. Rather, at that point I've been using the d,f,h partitions
for all my day-to-day work for 4 or 5 days, so I'm pretty confident that
they're ok. (If they didn't work ok, then I'd want to fix the problems
before proceeding.) Of course that "testing time" can be adjusted to
taste; it's probably more usefully measured in what-tasks-I've-done
rather than wall-clock-days.
> This would all be unnecessary if the bootloader could be informed that
--
-- "Jonathan Thornburg [remove -animal to reply]"
School of Mathematics, U of Southampton, England
"Space travel is utter bilge" -- common misquote of UK Astronomer Royal
Richard Woolley's remarks of 1956
"All this writing about space travel is utter bilge. To go to the
moon would cost as much as a major war." -- what he actually said
| Andy Whitcroft | clam |
| Tarkan Erimer | Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3 |
| Greg Kroah-Hartman | [PATCH 001/196] Chinese: Add the known_regression URI to the HOWTO |
| David Miller | Re: Slow DOWN, please!!! |
git: | |
| Arjan van de Ven | Re: [GIT]: Networking |
| Lennert Buytenhek | [PATCH 08/39] mv643xx_eth: nuke port status register bit defines |
| Jarek Poplawski | Re: HTB accuracy for high speed |
| Jarek Poplawski | [PATCH] pkt_sched: Destroy gen estimators under rtnl_lock(). |
