1 |
wabe wrote: |
2 |
> Dale <rdalek1967@×××××.com> wrote: |
3 |
> |
4 |
>> wabe wrote: |
5 |
>>> Hi folks, |
6 |
>>> |
7 |
>>> it seems that my system is broken after recompiling everything with |
8 |
>>> the new gcc5. |
9 |
>>> |
10 |
>>> I even cannot execute any command or start new programs. It is also |
11 |
>>> not possible to open a terminal or to login on console. |
12 |
>>> |
13 |
>>> When I type something in an already opened Xterm it says |
14 |
>>> "Ungültiger Maschinenbefehl" (in English probably something like |
15 |
>>> "invalid machine command"). |
16 |
>>> |
17 |
>>> On tty12 I can see messages like this: |
18 |
>>> |
19 |
>>> traps: ls[4055] trap invalid opcode ip:3cd3d91e024 sp:3d6a804abb0 |
20 |
>>> error:0 in libc-2.23.so[3cd3d865000+19d000] grsec: Illegal |
21 |
>>> instruction occured at 000003cd3d91e026 in /bin/ls[ls:4055] |
22 |
>>> uid/euid:0/0 gid/egid:0/0, parent... |
23 |
>>> |
24 |
>>> kernel is 4.8.17-hardened-r2 |
25 |
>>> |
26 |
>>> (uname -a is still working). |
27 |
>>> |
28 |
>>> I don't dare to reboot my system because I'm not sure that it will |
29 |
>>> boot. |
30 |
>>> |
31 |
>>> I don't even know if this email will reach the list, because claws |
32 |
>>> mail also dropped some errors about "bogofilter doesn't work". |
33 |
>>> |
34 |
>>> Any help would really appreciated. |
35 |
>>> |
36 |
>>> -- |
37 |
>>> Regards |
38 |
>>> wabe |
39 |
>>> |
40 |
>>> . |
41 |
>>> |
42 |
>> I wonder if this is related. When I did my switch, I hit a point |
43 |
>> where I had to start the alt sysreq key sequence. I only had to do |
44 |
>> the R and E part to get my console back tho. At the time, I thought |
45 |
>> it may just be some weird occurrence but now I wonder if it was |
46 |
>> related to some update. Once I got my console and login screen, I |
47 |
>> was able to login and carry on, after getting all the services back |
48 |
>> running that got killed off. I got the login screen back after |
49 |
>> hitting the E part. |
50 |
>> |
51 |
>> I wonder, will others run into this sort of problem? If so, what is |
52 |
>> the best way to get back to normal? For you, rebooting helped. For |
53 |
>> me, just starting the alt sysreq did the trick. It makes one wonder. |
54 |
> Hi Dale, |
55 |
> |
56 |
> thanks for this hint. Although I know about this "magic stuff", I don't |
57 |
> even have it enabled in my kernel. At the moment I'm happy that my |
58 |
> system is working again so that I can do some other urgent tasks. But |
59 |
> I will look closely at these function within the next time. Even when |
60 |
> it is not possible to "repair" the system in the case of a crash, it is |
61 |
> helpful to be able to write the cache to disk and reboot the system. |
62 |
> |
63 |
> -- |
64 |
> Regards |
65 |
> wabe |
66 |
> |
67 |
> |
68 |
|
69 |
|
70 |
It's rare that I have to go "all the way" with this but at least it does |
71 |
flush things through the system and properly unmounts file systems |
72 |
first. That alone is some comfort. I switched to ext4 which is |
73 |
supposed to handle those things pretty well BUT why risk it if you can |
74 |
unmount them properly, even if it is brute force. |
75 |
|
76 |
One thing about alt sysreq, once you get to a certain point, you have to |
77 |
go "all the way" with it. I've found that just the R and E generally |
78 |
gets things back to a point it is accessible and I don't have to go |
79 |
through it all. I think once you get to the I or S, it's all or nothing |
80 |
at that point. Even after the E, you may have to start/restart some |
81 |
init processes. FYI, I'm pretty sure the S is what syncs the file |
82 |
systems and the U unmounts everything. Of course, the B reboots. I |
83 |
can't recall what the others do exactly. |
84 |
|
85 |
I wonder if some guru that is familiar with this could shine some light |
86 |
and share some wisdom. |
87 |
|
88 |
Dale |
89 |
|
90 |
:-) :-) |