1 |
Am Sun, 19 Mar 2017 06:27:15 -0500 |
2 |
schrieb Dale <rdalek1967@×××××.com>: |
3 |
|
4 |
> Kai Krakow wrote: |
5 |
> > Am Sun, 19 Mar 2017 05:40:09 -0500 |
6 |
> > schrieb Dale <rdalek1967@×××××.com>: |
7 |
> > |
8 |
> >> Kai Krakow wrote: |
9 |
> [...] |
10 |
> [...] |
11 |
> >> [...] |
12 |
> [...] |
13 |
> [...] |
14 |
> >> I've likely had it happen on a regular console too. I have just |
15 |
> >> got so used to it, I don't pay it any attention. I suspect this |
16 |
> >> is a deep issue somewhere. Maybe even as low level as the kernel |
17 |
> >> somehow or close to it. |
18 |
> >> |
19 |
> >> It will be interesting to see what it is tho. Given how long it |
20 |
> >> has been doing it here at least, it's going to be a old |
21 |
> >> commit/change which may be difficult to track back. |
22 |
> > Well, I shouldn't say that probably, but some affected servers are |
23 |
> > still running on 3.0 or 3.2 kernels. Only the rest of the system was |
24 |
> > upgraded (mostly "glsa-check -f affected" only). So, I suspected an |
25 |
> > issue because of old kernel but new user space tools. |
26 |
> > |
27 |
> > But since some time ago my desktop machines are also affected (and |
28 |
> > those are almost bleeding edge, with ~amd64 gentoo-sources). |
29 |
> > |
30 |
> > So, more likely it's something in bash... Which is what I use. Which |
31 |
> > shell do you use? I could try using zsh tho I absolutely hate how it |
32 |
> > tries to be smarter about tab-completion and always steals trailing |
33 |
> > "/" away - which especially with rsync and mv can do some serious |
34 |
> > damage or at least unexpected results. |
35 |
> > |
36 |
> |
37 |
> |
38 |
> Here is mine: |
39 |
> |
40 |
> root@fireball / # uname -r |
41 |
> 4.5.2-gentoo |
42 |
> root@fireball / # |
43 |
> |
44 |
> As far as I know, I use bash. If you are talking about what I think |
45 |
> you are talking about. |
46 |
|
47 |
Yes, that's what I was talking about. |
48 |
|
49 |
Run ps, it should tell you the processes running in your current shell, |
50 |
including the shell itself: |
51 |
|
52 |
# ps |
53 |
PID TTY TIME CMD |
54 |
1256 pts/2 00:00:00 ps |
55 |
32059 pts/2 00:00:00 bash |
56 |
|
57 |
And you can see your default shell this way: |
58 |
|
59 |
# realpath /bin/sh |
60 |
/bin/dash |
61 |
|
62 |
Yes, dash for me, because it spawns much faster than bash, at least |
63 |
when running scripts. This can make a big difference with openrc. |
64 |
Meanwhile, I'm using systemd. |
65 |
|
66 |
> [IP-] [ ] app-shells/bash-4.3_p48-r1:0 |
67 |
|
68 |
Here, too: |
69 |
|
70 |
# equery list bash |
71 |
[IP-] [ ] app-shells/bash-4.3_p48-r1:0 |
72 |
|
73 |
> Given the age of your kernel, maybe it is above that level anyway. I |
74 |
> don't update my kernel often either. |
75 |
> |
76 |
> I'm going to be watching this thread tho. If I can share info which |
77 |
> may help narrow things down, I'll do that for sure. |
78 |
|
79 |
The problem is that this bug is totally non-deterministic... It fails |
80 |
once, next try it works as it should. |
81 |
|
82 |
If you can work out a way to reliably reproduce this bug, let me know. |
83 |
Then I'll try to work out what the problem is. |
84 |
|
85 |
-- |
86 |
Regards, |
87 |
Kai |
88 |
|
89 |
Replies to list-only preferred. |