| 1 |
Some thoughts: |
| 2 |
|
| 3 |
I recently did a stage 1 install and found that the process seems to |
| 4 |
have deteriorated to the point it was more work than it should have been |
| 5 |
- hence I see some of the reasons for abandoning it. In particular, the |
| 6 |
recompiling needed to bring it to a GCC 3.4.4 with all the options I |
| 7 |
needed meant that a stage 1 gained me nothing, and I lost quite a bit of |
| 8 |
time. |
| 9 |
|
| 10 |
The majority of systems I have recently installed have been tar over ssh |
| 11 |
from a running system (usually a LiveCD - I have P3/P4 and athlon - just |
| 12 |
choose the appropriate base). A small install can be up and running in |
| 13 |
less than 30 mins (IF you already have a running system!) - and its |
| 14 |
mostly preconfigured which is where I find I spend most of *MY* time. |
| 15 |
Only downside I have come across is cruft, but that can be managed. |
| 16 |
|
| 17 |
I consider this as the equivalent of a targeted (for my purposes) |
| 18 |
customised super stage3 install. |
| 19 |
|
| 20 |
With todays large hard disks, I also put aside a 4G reiserfs partition |
| 21 |
that contains a minimal install (inc a tailess /boot) to keep me working |
| 22 |
(i.e., the gateway has a basic webserver, squid, nat setup, mail |
| 23 |
server, ..., the desktop has fluxbox, OO and evolution - my main work |
| 24 |
tools and so on. Maintenance is done in a chroot, with an occasional |
| 25 |
test when scheduled with major kernel upgrades. If in fiddling, I have |
| 26 |
a disaster, I can keep working while rebuilding. If more than one |
| 27 |
physical HD is present, grub is installed in each MBR - many modern MB's |
| 28 |
allow you to choose which HD to boot from - quite handy! Worst comes to |
| 29 |
worst, a few minutes with tar and I have a basic, but fully configured |
| 30 |
base to start the recovery process back to the original system. |
| 31 |
|
| 32 |
I have found the 4G partition very handy when the raid array broke (disk |
| 33 |
failure - the 4G was in an unaffected area of the disk - non-raided, so |
| 34 |
was easily rescued), software problems (bad kernel upgrades) and just |
| 35 |
having the peace of mind that I can keep working through most disasters. |
| 36 |
I would highly recommend that this be a standard part of the install for |
| 37 |
critical systems (e.g., SOHO gateways), and especially for those who |
| 38 |
have only a single system to work with. |
| 39 |
|
| 40 |
With a little planning, it is possible to have an install once, and |
| 41 |
multiply/upgrade forever maintenance process - this is one of gentoo's |
| 42 |
current strengths. |
| 43 |
|
| 44 |
BillK |
| 45 |
|
| 46 |
On Tue, 2005-11-22 at 01:40 +0100, Holly Bostick wrote: |
| 47 |
> George Garvey schreef: |
| 48 |
> > On Mon, Nov 21, 2005 at 04:17:45PM +0100, Holly Bostick wrote: |
| 49 |
> > |
| 50 |
> >> reinstall, again I must wonder why he would complain that such a |
| 51 |
|
| 52 |
-- |
| 53 |
gentoo-user@g.o mailing list |
Archives