[LTP] [mtest01w] mtest01w test hangs

[LTP] [mtest01w] mtest01w test hangs

[SHASHANK BILGI]

[-- Attachment #1.1: Type: text/plain, Size: 1559 bytes --]

Dear all,

           I was running mtest01w test on a target board which has 80 MB or
RAM and 2GB of swap space. And this test keeps waiting forever, please find
the console logs below,

-8<---------------------------------------------------------
<<<test_start>>>
tag=mtest01w stime=2336
cmdline="mtest01 -p80 -w"
contacts=""
analysis=exit
<<<test_output>>>
[ 2351.111281] rmon:L
mtest01     0  TINFO  :  Total memory already used on system = 22796 kbytes
mtest01     0  TINFO  :  Total memory used needed to reach maximum =
1750246 kbytes
mtest01     0  TINFO  :  Filling up 80% of ram which is 1727450 kbytes
mtest01     1  TBROK  :  stopped at 716177408 bytes: errno=ENOMEM(12):
Cannot allocate memory


* When the malloc fails the child process seems to be calling tst_exit().
mtest01.c
.....
.....
        if (pid == 0) {
                bytecount = chunksize;
                while (1) {
                        if ((mem = malloc(chunksize)) == NULL) {
                                tst_resm(TBROK | TERRNO,
                                         "stopped at %lu bytes", bytecount);
                                free(pid_list);
                                tst_exit();
                        }

.....
.....

* but in this case i can see the child process in the process list(ps -ax)
in sleep state.

 333 ttyAM0   S      0:00 mtest01 -p80 -w
 334 ttyAM0   Z      0:14 [mtest01] <defunct>
 335 ttyAM0   S      0:13 mtest01 -p80 -w

Here is my query , why tst_exit() isn't ending the child process?
Somebody pls help!!

Regards ,
Shashank JB

[-- Attachment #1.2: Type: text/html, Size: 1917 bytes --]

[-- Attachment #2: Type: text/plain, Size: 394 bytes --]

------------------------------------------------------------------------------
See everything from the browser to the database with AppDynamics
Get end-to-end visibility with application monitoring from AppDynamics
Isolate bottlenecks and diagnose root cause in seconds.
Start your free trial of AppDynamics Pro today!
http://pubads.g.doubleclick.net/gampad/clk?id=48808831&iu=/4140/ostg.clktrk

[-- Attachment #3: Type: text/plain, Size: 155 bytes --]

_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] [mtest01w] mtest01w test hangs

[chrubis]

Hi!
> <<<test_start>>>
> tag=mtest01w stime=2336
> cmdline="mtest01 -p80 -w"
> contacts=""
> analysis=exit
> <<<test_output>>>
> [ 2351.111281] rmon:L
> mtest01     0  TINFO  :  Total memory already used on system = 22796 kbytes
> mtest01     0  TINFO  :  Total memory used needed to reach maximum =
> 1750246 kbytes
> mtest01     0  TINFO  :  Filling up 80% of ram which is 1727450 kbytes
> mtest01     1  TBROK  :  stopped at 716177408 bytes: errno=ENOMEM(12):
> Cannot allocate memory
> 
> 
> * When the malloc fails the child process seems to be calling tst_exit().
> mtest01.c
> .....
> .....
>         if (pid == 0) {
>                 bytecount = chunksize;
>                 while (1) {
>                         if ((mem = malloc(chunksize)) == NULL) {
>                                 tst_resm(TBROK | TERRNO,
>                                          "stopped at %lu bytes", bytecount);
>                                 free(pid_list);
>                                 tst_exit();
>                         }
> 
> .....
> .....
> 
> * but in this case i can see the child process in the process list(ps -ax)
> in sleep state.
> 
>  333 ttyAM0   S      0:00 mtest01 -p80 -w
>  334 ttyAM0   Z      0:14 [mtest01] <defunct>
>  335 ttyAM0   S      0:13 mtest01 -p80 -w
> 
> Here is my query , why tst_exit() isn't ending the child process?

Using tst_exit() and any other tst interface from child is a mistake but
tst_exit() should end the child process.

However there may be other childrens still running as tst_exit() is just
equivalent to exit() (it does flush output buffers and exits the process
with result as return value but nothing more).

So my guess is that the rest of the childrens continue to fill the
chunks and then waits to be reaped by the parent. Which may hang the
machine if there is not enough memory.

-- 
Cyril Hrubis
chrubis@suse.cz

------------------------------------------------------------------------------
See everything from the browser to the database with AppDynamics
Get end-to-end visibility with application monitoring from AppDynamics
Isolate bottlenecks and diagnose root cause in seconds.
Start your free trial of AppDynamics Pro today!
http://pubads.g.doubleclick.net/gampad/clk?id=48808831&iu=/4140/ostg.clktrk
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] [mtest01w] mtest01w test hangs

[chrubis]

Hi!
> I have one more query, when i run mtest01w on my target, it gives a message
> "stopped at 731906048 bytes: errno=ENOMEM(12): Cannot allocate memory".
> 
> Should this be considered as a test case failure, If not, why?

That is hard to say without further analysis.

Generally malloc() is allowed to return ENOMEM when machine memory was
used up.

But conditions under this happens depends on ulimits, memory overcommit
settings, size and availability of swap etc.

-- 
Cyril Hrubis
chrubis@suse.cz

------------------------------------------------------------------------------
See everything from the browser to the database with AppDynamics
Get end-to-end visibility with application monitoring from AppDynamics
Isolate bottlenecks and diagnose root cause in seconds.
Start your free trial of AppDynamics Pro today!
http://pubads.g.doubleclick.net/gampad/clk?id=48808831&iu=/4140/ostg.clktrk
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] [mtest01w] mtest01w test hangs

[chrubis]

Hi!
> In my case 2 child processes from mtest01w try to allocate 1GB each and
> third one allocates around 100 MB.
> Both of the first 2 child processes throws this error "errno=ENOMEM(12):
> Cannot allocate memory" around 700MB. And third successfully finishes its
> job.
> Here are some details.
> * Actual physical memory: 80MB
> * Swap space is 2GB
> * /proc/sys/vm/overcommit_memory is 1
> * /proc/sys/vm/overcommit_ratio is 50
> 
> Can somebody please tell me why this might be happening?

Hmm, are you aware that in this case you RAM + SWAP is lesser than the
memory the program is trying to allocate, which will not work no matter
what.

-- 
Cyril Hrubis
chrubis@suse.cz

------------------------------------------------------------------------------
Get your SQL database under version control now!
Version control is standard for application code, but databases havent 
caught up. So what steps can you take to put your SQL databases under 
version control? Why should you start doing it? Read more to find out.
http://pubads.g.doubleclick.net/gampad/clk?id=49501711&iu=/4140/ostg.clktrk
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list