1 |
2016.Június 2.(Cs) 21:39 időpontban "Tóth Attila" ezt írta: |
2 |
> 2016.Június 2.(Cs) 02:31 időpontban Max R.D. Parmer ezt írta: |
3 |
>> Not necessarily the ideal solution, but have you tried twiddling with |
4 |
>> the stack size in limits.conf? |
5 |
> |
6 |
> I checked an the system-wide defaults apply to systemd, which is: 8M soft |
7 |
> limit and _unlimited_ hard limit for stack size. So after exceeding soft |
8 |
> limit systemd segfaults and tries to dump core. |
9 |
> |
10 |
> cat /proc/1/limits |
11 |
> Limit Soft Limit Hard Limit Units |
12 |
> Max stack size 8388608 unlimited bytes |
13 |
> |
14 |
> I expect any process to handle a situation of trying to exceed soft limit |
15 |
> with unlimited hard limit in another way than segfaulting and attempting |
16 |
> to dump core... |
17 |
|
18 |
Increasing the limit doesn't fix the issue - I'm not surprised about that. |
19 |
|
20 |
For those who are not familiar: systemd doesn't respect limits.conf. In |
21 |
system.conf the default values can be configured and per unit limits can |
22 |
be specified. To my surprise, systemd doesn't seem to pay attention to |
23 |
it's own configuration file. In order to provide increased stack limit for |
24 |
init, I also modified the kernel defaults. With no success. |
25 |
|
26 |
>> If I read this right, grsec limits the size of the stack, which causes |
27 |
>> the process to segfault. |
28 |
> |
29 |
> I think grsec does not enforce any stack limits here, it just reports the |
30 |
> issue and makes it more visible. |
31 |
|
32 |
I did a bisect and it turns out a this commit is responsible for the |
33 |
symptoms: |
34 |
https://github.com/systemd/systemd/commit/d054f0a4d451120c26494263fc4dc175bfd405b1 |
35 |
tree-wide: use xsprintf() where applicable |
36 |
|
37 |
I try to contact the developer. Whether he has an idea on what is |
38 |
happening here. |
39 |
|
40 |
BR: Dw. |
41 |
-- |
42 |
dr Tóth Attila, Radiológus, 06-20-825-8057 |
43 |
Attila Toth MD, Radiologist, +36-20-825-8057 |