[Pacemaker] [Problem]When Pacemaker uses a new version of glib, g_source_remove fails.

renayama19661014 at ybb.ne.jp renayama19661014 at ybb.ne.jp
Fri Oct 10 01:16:15 UTC 2014


Hi Andrew,

Setting of gdb of the Ubuntu environment does not yet go well and I touch lrmd and cannot acquire trace.
Please wait for this a little more.


But.. I let lrmd terminate abnormally when g_source_remove() of cancel_recurring_action() returned FALSE.
-----
gboolean
cancel_recurring_action(svc_action_t * op)
{
    crm_info("Cancelling operation %s", op->id);

    if (recurring_actions) {
        g_hash_table_remove(recurring_actions, op->id);
    }

    if (op->opaque->repeat_timer) {
        if (g_source_remove(op->opaque->repeat_timer) == FALSE)  {
                abort();
        }
(snip)
-------core----
#0  0x00007f30aa60ff79 in __GI_raise (sig=sig at entry=6) at ../nptl/sysdeps/unix/sysv/linux/raise.c:56

56      ../nptl/sysdeps/unix/sysv/linux/raise.c: No such file or directory.
(gdb) where
#0  0x00007f30aa60ff79 in __GI_raise (sig=sig at entry=6) at ../nptl/sysdeps/unix/sysv/linux/raise.c:56
#1  0x00007f30aa613388 in __GI_abort () at abort.c:89
#2  0x00007f30aadcde77 in crm_abort (file=file at entry=0x7f30aae0152b "logging.c", 
    function=function at entry=0x7f30aae028c0 <__FUNCTION__.23262> "crm_glib_handler", line=line at entry=73, 
    assert_condition=assert_condition at entry=0x19d2ad0 "Source ID 63 was not found when attempting to remove it", do_core=do_core at entry=1, 
    do_fork=<optimized out>, do_fork at entry=1) at utils.c:1195
#3  0x00007f30aadf5ca7 in crm_glib_handler (log_domain=0x7f30aa35eb6e "GLib", flags=<optimized out>, 
    message=0x19d2ad0 "Source ID 63 was not found when attempting to remove it", user_data=<optimized out>) at logging.c:73
#4  0x00007f30aa320ae1 in g_logv () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#5  0x00007f30aa320d72 in g_log () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#6  0x00007f30aa318c5c in g_source_remove () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#7  0x00007f30aabb2b55 in cancel_recurring_action (op=op at entry=0x19caa90) at services.c:363
#8  0x00007f30aabb2bee in services_action_cancel (name=name at entry=0x19d0530 "dummy3", action=<optimized out>, interval=interval at entry=10000)
    at services.c:385
#9  0x000000000040405a in cancel_op (rsc_id=rsc_id at entry=0x19d0530 "dummy3", action=action at entry=0x19cec10 "monitor", interval=10000)
    at lrmd.c:1404
#10 0x000000000040614f in process_lrmd_rsc_cancel (client=0x19c8290, id=74, request=0x19ca8a0) at lrmd.c:1468
#11 process_lrmd_message (client=client at entry=0x19c8290, id=74, request=request at entry=0x19ca8a0) at lrmd.c:1507
#12 0x0000000000402bac in lrmd_ipc_dispatch (c=0x19c79c0, data=<optimized out>, size=361) at main.c:148
#13 0x00007f30aa07b4d9 in qb_ipcs_dispatch_connection_request () from /usr/lib/libqb.so.0
#14 0x00007f30aadf209d in gio_read_socket (gio=<optimized out>, condition=G_IO_IN, data=0x19c68a8) at mainloop.c:437
#15 0x00007f30aa319ce5 in g_main_context_dispatch () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
---Type <return> to continue, or q <return> to quit---
#16 0x00007f30aa31a048 in ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#17 0x00007f30aa31a30a in g_main_loop_run () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#18 0x0000000000402774 in main (argc=<optimized out>, argv=0x7fffcdd90b88) at main.c:344
---------

Best Regards,
Hideo Yamauchi.



----- Original Message -----
> From: "renayama19661014 at ybb.ne.jp" <renayama19661014 at ybb.ne.jp>
> To: Andrew Beekhof <andrew at beekhof.net>
> Cc: The Pacemaker cluster resource manager <pacemaker at oss.clusterlabs.org>
> Date: 2014/10/7, Tue 11:15
> Subject: Re: [Pacemaker] [Problem]When Pacemaker uses a new version of glib, g_source_remove fails.
> 
> Hi Andrew,
> 
>>  Not quite. Returning FALSE from the callback also removes the source from 
> glib.
>>  So your test case effectively removes t1 twice: once implicitly by 
> returning 
>>  FALSE in timer_func1() and then again explicitly in timer_func3()
> 
> 
> Your opinion is right.
> 
> 
> If Pacemaker repeats and does not remove the resources which timer concluded in 
> FALSE, glib does not return the error.
> 
> 
> Many Thanks,
> Hideo Yamauchi.
> 
> 
> ----- Original Message -----
>>  From: Andrew Beekhof <andrew at beekhof.net>
>>  To: renayama19661014 at ybb.ne.jp
>>  Cc: The Pacemaker cluster resource manager 
> <pacemaker at oss.clusterlabs.org>
>>  Date: 2014/10/7, Tue 11:06
>>  Subject: Re: [Pacemaker] [Problem]When Pacemaker uses a new version of 
> glib, g_source_remove fails.
>> 
>> 
>>  On 7 Oct 2014, at 1:03 pm, renayama19661014 at ybb.ne.jp wrote:
>> 
>>>   Hi Andrew,
>>> 
>>>>>   These problems seem to be due to a correction of next glib 
> somehow 
>>  or 
>>>>   other.
>>>>>     * 
>>>> 
>> 
> https://github.com/GNOME/glib/commit/393503ba5bdc7c09cd46b716aaf3d2c63a6c7f9c
>>>>    
>>>>   The glib behaviour on unbuntu seems reasonable, removing a source 
>>  multiple times 
>>>>   IS a valid error.
>>>>   I need the stack trace to know where/how this situation can occur 
> in 
>>  pacemaker.
>>> 
>>> 
>>>   Pacemaker does not remove resources several times as far as I 
> confirmed it.
>>>   In Ubuntu(glib2.40), an error occurs just to remove resources first.
>> 
>>  Not quite. Returning FALSE from the callback also removes the source from 
> glib.
>>  So your test case effectively removes t1 twice: once implicitly by 
> returning 
>>  FALSE in timer_func1() and then again explicitly in timer_func3()
>> 
>>> 
>>>   Confirmation and the deletion of resources seem to be necessary not to 
> 
>>  produce an error in Ubuntu.
>>>   And this works well in glib of RHEL6.x.(and RHEL7.0)
>>> 
>>>           if (g_main_context_find_source_by_id (NULL, t1) != NULL) {
>>>                   g_source_remove(t1);
>>>           }
>>> 
>>>   I send it to you after acquiring stack trace.
>>> 
>>>   Many Thanks!
>>>   Hideo Yamauchi.
>>> 
>>>   ----- Original Message -----
>>>>   From: Andrew Beekhof <andrew at beekhof.net>
>>>>   To: renayama19661014 at ybb.ne.jp; The Pacemaker cluster resource 
> manager 
>>  <pacemaker at oss.clusterlabs.org>
>>>>   Cc: 
>>>>   Date: 2014/10/7, Tue 09:44
>>>>   Subject: Re: [Pacemaker] [Problem]When Pacemaker uses a new 
> version of 
>>  glib, g_source_remove fails.
>>>> 
>>>> 
>>>>   On 6 Oct 2014, at 4:09 pm, renayama19661014 at ybb.ne.jp wrote:
>>>> 
>>>>>   Hi All,
>>>>> 
>>>>>   When I move the next sample in RHEL6.5(glib2-2.22.5-7.el6) and 
> 
>>>>   Ubuntu14.04(libglib2.0-0:amd64 2.40.0-2), movement is different.
>>>>> 
>>>>>     * Sample : test2.c
>>>>>   {{{
>>>>>   #include <stdio.h>
>>>>>   #include <stdlib.h>
>>>>>   #include <glib.h>
>>>>>   #include <sys/times.h>
>>>>>   guint t1, t2, t3;
>>>>>   gboolean timer_func2(gpointer data){
>>>>>            printf("TIMER EXPIRE!2\n");
>>>>>            fflush(stdout);
>>>>>            return FALSE;
>>>>>   }
>>>>>   gboolean timer_func1(gpointer data){
>>>>>            clock_t         ret;
>>>>>            struct tms buff;
>>>>> 
>>>>>            ret = times(&buff);
>>>>>            printf("TIMER EXPIRE!1 %d\n", 
> (int)ret);
>>>>>            fflush(stdout);
>>>>>            return FALSE;
>>>>>   }
>>>>>   gboolean timer_func3(gpointer data){
>>>>>            printf("TIMER EXPIRE 3!\n");
>>>>>            fflush(stdout);
>>>>>            printf("remove timer1!\n");
>>>>> 
>>>>>            fflush(stdout);
>>>>>            g_source_remove(t1);
>>>>>            printf("remove timer2!\n");
>>>>>            fflush(stdout);
>>>>>            g_source_remove(t2);
>>>>>            printf("remove timer3!\n");
>>>>>            fflush(stdout);
>>>>>            g_source_remove(t3);
>>>>>            return FALSE;
>>>>>   }
>>>>>   int main(int argc, char** argv){
>>>>>            GMainLoop *m;
>>>>>            clock_t         ret;
>>>>>            struct tms buff;
>>>>>            gint64 t;
>>>>>            m = g_main_new(FALSE);
>>>>>            t1 = g_timeout_add(1000, timer_func1, NULL);
>>>>>            t2 = g_timeout_add(60000, timer_func2, NULL);
>>>>>            t3 = g_timeout_add(5000, timer_func3, NULL);
>>>>>            ret = times(&buff);
>>>>>            printf("START! %d\n", (int)ret);
>>>>>            g_main_run(m);
>>>>>   }
>>>>> 
>>>>>   }}}
>>>>>     * Result
>>>>>   ---- RHEL6.5(glib2-2.22.5-7.el6) ---- 
>>>>>   [root at snmp1 ~]# ./test2
>>>>>   START! 429576012
>>>>>   TIMER EXPIRE!1 429576112
>>>>>   TIMER EXPIRE 3!
>>>>>   remove timer1!
>>>>>   remove timer2!
>>>>>   remove timer3!
>>>>> 
>>>>>   ---- Ubuntu14.04(libglib2.0-0:amd64 2.40.0-2) ----
>>>>>   root at a1be102:~# ./test2
>>>>>   START! 1718163089
>>>>>   TIMER EXPIRE!1 1718163189
>>>>>   TIMER EXPIRE 3!
>>>>>   remove timer1!
>>>>> 
>>>>>   (process:1410): GLib-CRITICAL **: Source ID 1 was not found 
> when 
>>  attempting 
>>>>   to remove it
>>>>>   remove timer2!
>>>>>   remove timer3!
>>>>> 
>>>>> 
>>>>>   These problems seem to be due to a correction of next glib 
> somehow 
>>  or 
>>>>   other.
>>>>>     * 
>>>> 
>> 
> https://github.com/GNOME/glib/commit/393503ba5bdc7c09cd46b716aaf3d2c63a6c7f9c
>>>> 
>>>>   The glib behaviour on unbuntu seems reasonable, removing a source 
>>  multiple times 
>>>>   IS a valid error.
>>>>   I need the stack trace to know where/how this situation can occur 
> in 
>>  pacemaker.
>>>> 
>>>>> 
>>>>>   In g_source_remove() until before change, the deletion of the 
> timer 
>>  which 
>>>>   practice completed is possible, but g_source_remove() after the 
> change 
>>  causes an 
>>>>   error.
>>>>> 
>>>>>   Under this influence, we get the following crit error in the 
>>  environment of 
>>>>   Pacemaker using a new version of glib.
>>>>> 
>>>>>   lrmd[1632]:    error: crm_abort: crm_glib_handler: Forked 
> child 
>>  1840 to 
>>>>>   record non-fatal assert at logging.c:73 : Source ID 51 was not 
> 
>>  found when 
>>>>>   attempting to remove it
>>>>>   lrmd[1632]:    crit: crm_glib_handler: GLib: Source ID 51 was 
> not 
>>  found 
>>>>>   when attempting to remove it
>>>>> 
>>>>>   It seems that some kind of coping is necessary in Pacemaker 
> when I 
>>  think 
>>>>   about next.
>>>>>     * Distribution using a new version of glib including Ubuntu.
>>>>>     * Version up of future glib of RHEL.
>>>>> 
>>>>>   A similar problem is reported in the ML.
>>>>>     * 
>>  http://www.gossamer-threads.com/lists/linuxha/pacemaker/91333#91333
>>>>>     * 
> http://www.gossamer-threads.com/lists/linuxha/pacemaker/92408
>>>>> 
>>>>>   Best Regards,
>>>>>   Hideo Yamauchi.
>>>>> 
>>>>>   _______________________________________________
>>>>>   Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
>>>>>   http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>>>>> 
>>>>>   Project Home: http://www.clusterlabs.org
>>>>>   Getting started: 
>>  http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
>>>>>   Bugs: http://bugs.clusterlabs.org
>>>> 
>> 
> 
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
> 
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org
> 




More information about the Pacemaker mailing list