1 |
Beso <givemesugarr@×××××.com> posted |
2 |
d257c3560901221141g57061763y2cdb773058920241@××××××××××.com, excerpted |
3 |
below, on Thu, 22 Jan 2009 20:41:18 +0100: |
4 |
|
5 |
>> Not that anyone else necessarily needs to use my "everything portage |
6 |
>> touches on one partition" strategy, but I certainly learned /my/ |
7 |
>> lesson, and don't intend on screwing /that/ one up here again. It's |
8 |
>> worth considering, anyway. YMMV. |
9 |
>> |
10 |
> i really don't really understand how you could have had this issue if |
11 |
> you'd mount the |
12 |
> lvm partition at boot via fstab. it's most likely to not happen |
13 |
> anything. |
14 |
|
15 |
It seems I need to explain a bit further... |
16 |
|
17 |
My triggering problem was that my A/C died, in Phoenix, AZ, in the |
18 |
summer, where it's not unusual to see 45C/115F in the shade, but the |
19 |
computer was left running in the hot house where the ambient air temp |
20 |
likely hit 60C/140F, and the still-spinning disks would have easily hit |
21 |
70C and may have been as high as 85C! |
22 |
|
23 |
What happened was that under those conditions, the platter expanded such |
24 |
that the heads were fully crashed, scratching grooves into it. Wherever |
25 |
the heads moved while the disk was overheated like that, was very quickly |
26 |
an unreadable mess! |
27 |
|
28 |
Well, as alluded to previously, I had backup partitions of most |
29 |
everything. They weren't mounted at the time and weren't read/written, |
30 |
so didn't get grooved out and survived pretty much intact. |
31 |
|
32 |
The problem was that I ended up with a mix of some "current" partitions |
33 |
and some backup partitions and a package database that wasn't at all in |
34 |
sync with what was actually on disk. What was on disk was still |
35 |
operational, and I actually ran the disks for some time (after getting a |
36 |
new A/C and cooling everything off, obviously) without a problem with |
37 |
further data loss, but as I said previously, I learned my lesson. Now, |
38 |
everything portage touches including the database it keeps track of it |
39 |
all in is on the same partition, so if for whatever reason I have to |
40 |
revert to the backups, the older snapshot is consistently the same age, |
41 |
and the package database remains in sync with what's actually running. |
42 |
|
43 |
-- |
44 |
Duncan - List replies preferred. No HTML msgs. |
45 |
"Every nonfree program has a lord, a master -- |
46 |
and if you use the program, he is your master." Richard Stallman |