1 |
On Thursday 19 Nov 2015 19:17:11 Alan McKinnon wrote: |
2 |
|
3 |
> Does it fail consistently? IOW, now that everything else has settled |
4 |
> down, does "/etc/init.d/chrony start" still fail? |
5 |
|
6 |
Yes, it's a hard fault: |
7 |
|
8 |
# /etc/init.d/chronyd start |
9 |
* Caching service dependencies ... [ ok ] |
10 |
* Starting chronyd ... |
11 |
* start-stop-daemon: caught an interrupt |
12 |
* start-stop-daemon: /usr/sbin/chronyd died |
13 |
* Failed to start chronyd [ !! ] |
14 |
* ERROR: chronyd failed to start |
15 |
|
16 |
> What about logs? |
17 |
|
18 |
None evident. |
19 |
|
20 |
> If none, you can inspect the start-stop-daemon line in the init file, |
21 |
> get the command line it launches chrony with, and see what that prints |
22 |
> to the console. |
23 |
|
24 |
Here's where it gets strange. When I run that command, chronyd runs hunky- |
25 |
dory, but calling it from /etc/init.d/chronyd gets me the error above. This |
26 |
is what works: |
27 |
|
28 |
# start-stop-daemon --start --exec /usr/sbin/chronyd --pidfile |
29 |
/run/chronyd.pid -- -f /etc/chrony/chrony.conf -s -r |
30 |
|
31 |
And this stops it: |
32 |
|
33 |
# start-stop-daemon --stop --pidfile /run/chronyd.pid |
34 |
|
35 |
All those file names are copied directly from the init.d file. I did wonder |
36 |
about a timing problem somewhere, but I wouldn't expect that always to turn |
37 |
out the same way. |
38 |
|
39 |
As I said, chronyd works on this four-core i5 box, just not on the two-core |
40 |
Atom. |
41 |
|
42 |
-- |
43 |
Rgds |
44 |
Peter |