[CRIU] Parasite code is not executed

Pavel Emelyanov xemul at parallels.com
Wed Mar 18 06:41:23 PDT 2015


On 03/18/2015 04:24 PM, Vitaliy Kharin wrote:
> criu 1.5

OK. Let's then try to look at the dumpee. There's a file /proc/$pid/stack,
can you check its contents for the dumpee process, where does _it_ hangs?

And one more question -- how easy is it to reproduce this situation?

Thanks,
Pavel

> 2015-03-18 16:21 GMT+03:00 Pavel Emelyanov <xemul at parallels.com>:
>> On 03/18/2015 01:04 PM, Vitaliy Kharin wrote:
>>> Hi all!
>>> I am using
>>> Linux 3.16.0-4-amd64 #1 SMP Debian 3.16.7-ckt7-1 (2015-03-01) x86_64 GNU/Linux
>>
>> And which version of CRIU do you use?
>>
>>> sudo ./criu check --ms
>>> Warn  (cr-check.c:188): Skipping unssuported PR_SET_MM_MAP
>>> Warn  (tun.c:58): Skipping tun support check
>>> Warn  (timerfd.c:44): timerfd: Skipping timerfd support check
>>> Warn  (cr-check.c:296): Skipping mnt_id support check
>>> Warn  (cr-check.c:668): Skipping unsupported AIO remap
>>> Looks good.
>>>
>>> A tried to dump my app:
>>> sudo criu dump -t PID -vvv -o dump.log && echo OK
>>>
>>> Process PID stopped but exists, 'OK' is not printed
>>>
>>> Tail of dump.log contains:
>>> (00.040686) Putting parasite blob into 0x7f4451fd2000->0x7f087a105000
>>> (00.040705) Dumping GP/FPU registers for 10935
>>> (00.040713) Putting tsock into pid 10935
>>>
>>> I found out that criu is blocked on an accept call:
>>> static int accept_tsock(struct parasite_ctl *ctl)
>>> ...
>>> sock = accept(ask, NULL, 0);
>>>
>>> Seems to be parasite code inside of dumpee  did not create connection.
>>> Could you assist me to find out the cause?
> 
> 



More information about the CRIU mailing list