1 |
On Sat, 25 Apr 2015 21:32:27 +0000 (UTC) |
2 |
james <wireless@×××××××××××.com> wrote: |
3 |
|
4 |
> »Q« <boxcars <at> gmx.net> writes: |
5 |
> |
6 |
> > On my amd64 laptop, on resume from hibernation, wicd almost always |
7 |
> > fails to restart. This problem goes back to when I set up the |
8 |
> > machine, over a year ago, but it seems to have gotten a lot worse |
9 |
> > in the last month or so. |
10 |
> |
11 |
> Shot in the dark: Check (and google for issues with obscure bios |
12 |
> settings for your version of the bios. Sometime a bios |
13 |
> upgrade|downgrade can help, but my understanding is that hibernate on |
14 |
> laptops is so varied that often some brands do not work even with |
15 |
> redmond products. |
16 |
> |
17 |
> some marginally useful keywords: PXE and (u)efi |
18 |
|
19 |
I was hoping someone would hand me a magic bullet so I wouldn't have to |
20 |
go fishing for that stuff, heh. Your advice didn't help me directly, |
21 |
but I think just seeing it in writing was enough to prompt me to some |
22 |
better thinking than I was doing before, so thanks! |
23 |
|
24 |
I procrastinated about the search until my brain cell fired and I tried |
25 |
`hibernate --no-suspend` a few times and found that wicd didn't restart |
26 |
even when the machine didn't actually hibernate. Then with wicd |
27 |
running, I tried just `/etc/init.d/wicd restart` -- every time, it |
28 |
stopped but failed to restart. But if I waited a few seconds after |
29 |
stopping wicd, then it would start. It was a great relief to have |
30 |
isolated the problem from the hibernate process itself. |
31 |
|
32 |
I've solved it by abandoning hibernate's RestartServices directive and |
33 |
instead stopping and starting wicd more explicitly with |
34 |
|
35 |
OnSuspend 10 /etc/init.d/wicd stop |
36 |
OnResume 10 /usr/local/bin/wicd-delayed-start |
37 |
|
38 |
Where wicd-delayed-start is a script with |
39 |
|
40 |
/usr/bin/sleep 3 && /etc/init.d/wicd start |
41 |
|
42 |
I have no idea what's at the root of the apparent timing issue -- I |
43 |
figure it's that the networking gremlins refuse to be called back to |
44 |
work too soon once they've been told they can take a break. |
45 |
|
46 |
In case it helps future searchers, I think the relevant hardware info |
47 |
is: Lenovo IdeaPad y510p with Intel Corporation Centrino Wireless-N |
48 |
2230 (rev c4). |