[Pacemaker] Among monitor of STONITH resource, monitor of other resources is not performed

David Vossel dvossel at redhat.com
Tue Jul 31 15:11:44 UTC 2012


----- Original Message -----
> From: "Kazunori INOUE" <inouekazu at intellilink.co.jp>
> To: "pacemaker at oss" <pacemaker at oss.clusterlabs.org>
> Cc: shimazakik at intellilink.co.jp
> Sent: Monday, July 30, 2012 5:15:26 AM
> Subject: [Pacemaker] Among monitor of STONITH resource, monitor of other resources is not performed
> 
> Hi,
> 
> I am using Pacemaker-1.1.
> - glue       (2012 Jul 16) 2719:18489f275f75
> - libqb      (2012 Jul 19) 11b20e19beff7f1b6003be0b4c73da8ecf936442
> - corosync   (2012 Jul 19) b9eb19e623d2b69c86cf78f1fa50a004c804ac20
> - pacemaker  (2012 Jul 29) 33119da31c235710195c783e5c9a32c6e95b3efc
> 
> The monitor operation of other resources is not performed during the
> monitor of STONITH resource.

Looking at the code I can see this is true.  How important is it for the stonith monitors to be able to be performed in parallel with the other resource monitors?   I would expect the stonith monitor interval to be quite large compared to all the other monitors, meaning that it only executes once every hour or so.

-- Vossel

> 1. service corosync start ; service pacemaker start
> 2. cibadmin -U -x test.xml
>    (STONITH resource + Dummy resource are started on the same node)
> 3. After libvirt (STONITH resource) started, sleep was added to
> status()
>    of libvirt.
> 
>    [root at dev1 external]# diff -u libvirt.ORG libvirt
>    --- libvirt.ORG 2012-07-17 13:10:01.000000000 +0900
>    +++ libvirt     2012-07-30 13:36:19.661431208 +0900
>    @@ -221,6 +221,7 @@
>         ;;
> 
>         status)
>    +    sleep 3600
>         libvirt_check_config
>         libvirt_status
>         exit $?
> 
>    [root at dev1 ~]# ps -ef|egrep
>    "UID|corosync|pacemaker|stonith|fence|sleep"
>    UID    PID  PPID  C STIME TTY     TIME CMD
>    root 18567     1  0 18:47 ?   00:00:02 corosync
>    root 18585     1  0 18:47 ?   00:00:00 pacemakerd
>    496  18587 18585  0 18:47 ?   00:00:00 /usr/libexec/pacemaker/cib
>    root 18588 18585  0 18:47 ?   00:00:00
>    /usr/libexec/pacemaker/stonithd
>    root 18589 18585 76 18:47 ?   00:05:27 /usr/libexec/pacemaker/lrmd
>    496  18590 18585  0 18:47 ?   00:00:00
>    /usr/libexec/pacemaker/attrd
>    496  18591 18585  0 18:47 ?   00:00:00
>    /usr/libexec/pacemaker/pengine
>    496  18592 18585  0 18:47 ?   00:00:00 /usr/libexec/pacemaker/crmd
>    root 18767 18588  0 18:48 ?   00:00:00 /usr/bin/perl
>    /usr/sbin/fence_legacy
>    root 18768 18767  0 18:48 ?   00:00:00 stonith -t external/libvirt
>    -E -S
>    root 18778 18768  0 18:48 ?   00:00:00 /bin/sh
>    /usr/lib64/stonith/plugins/external/libvirt status
>    root 18792 18778  0 18:48 ?   00:00:00 sleep 3600
> 
> 4. Then monitor of Dummy resource is not performed.
> 
> The following is behavior of lrmd at that time.
> 
> # gdb /usr/libexec/pacemaker/lrmd `pgrep lrmd`
> (gdb) bt
> #0  0x0000003f808e83e2 in recv () from /lib64/libc.so.6
> #1  0x00007f0de3820062 in qb_ipc_us_recv_at_most (one_way=0x1118ee8,
> msg=0x111f390,
>     len=20480, timeout=500) at ipc_us.c:299
> #2  0x00007f0de381a28e in qb_ipcc_recv (c=0x1118ba0,
> msg_ptr=0x111f390, msg_len=
>     20480, ms_timeout=500) at ipcc.c:249
> #3  0x00007f0de42bc5fb in crm_ipc_send (client=0x111b580,
> message=0x111d860, reply=
>     0x7fffc8ec4c60, ms_timeout=61060000) at ipc.c:517
> #4  0x00007f0de3c97e29 in stonith_send_command (stonith=0x111a6c0,
> op=
>     0x7f0de3c998dd "st_execute", data=0x1119170, output_data=0x0,
>     call_options=4096,
>     timeout=61000) at st_client.c:1676
> #5  0x00007f0de3c94bd1 in stonith_api_call (stonith=0x111a6c0,
> call_options=4096,
>     id=0x1129860 "f-2", action=0x7f0de3c998f7 "monitor", victim=0x0,
>     timeout=61000,
>     output=0x0) at st_client.c:951
> #6  0x00007f0de3c94d31 in stonith_api_monitor (stonith=0x111a6c0,
> call_options=4096,
>     id=0x1129860 "f-2", timeout=61000) at st_client.c:985
> #7  0x00000000004044e2 in lrmd_rsc_execute_stonith (rsc=0x111f130,
> cmd=0x1129660) at
>     lrmd.c:522
> #8  0x0000000000404cd6 in lrmd_rsc_execute (rsc=0x111f130) at
> lrmd.c:667
> #9  0x0000000000404d2d in lrmd_rsc_dispatch (user_data=0x111f130) at
> lrmd.c:678
> #10 0x00007f0de42dcd00 in crm_trigger_dispatch (source=0x111f300,
> callback=
>     0x404d06 <lrmd_rsc_dispatch>, userdata=0x111f300) at
>     mainloop.c:105
> #11 0x0000003642638f0e in g_main_context_dispatch () from
> /lib64/libglib-2.0.so.0
> #12 0x000000364263c938 in ?? () from /lib64/libglib-2.0.so.0
> #13 0x000000364263cd55 in g_main_loop_run () from
> /lib64/libglib-2.0.so.0
> #14 0x0000000000402d3f in main (argc=1, argv=0x7fffc8ec5188) at
> main.c:285
> (gdb) fin
> Run till exit from #0  0x0000003f808e83e2 in recv () from
> /lib64/libc.so.6
> 0x00007f0de3820062 in qb_ipc_us_recv_at_most (one_way=0x1118ee8,
> msg=0x111f390,
>     len=20480, timeout=500) at ipc_us.c:299
> 299         result = recv(one_way->u.us.sock, &data[processed],
> to_recv,
> (gdb) fin
> Run till exit from #0  0x00007f0de3820062 in qb_ipc_us_recv_at_most
> (one_way=
>     0x1118ee8, msg=0x111f390, len=20480, timeout=500) at ipc_us.c:299
> 0x00007f0de381a28e in qb_ipcc_recv (c=0x1118ba0, msg_ptr=0x111f390,
> msg_len=20480,
>     ms_timeout=500) at ipcc.c:249
> 249         res = c->funcs.recv(&c->response, msg_ptr, msg_len,
> ms_timeout);
> Value returned is $1 = -11
> (gdb) fin
> Run till exit from #0  0x00007f0de381a28e in qb_ipcc_recv
> (c=0x1118ba0, msg_ptr=
>     0x111f390, msg_len=20480, ms_timeout=500) at ipcc.c:249
> 0x00007f0de42bc5fb in crm_ipc_send (client=0x111b580,
> message=0x111d860, reply=
>     0x7fffc8ec4c60, ms_timeout=61060000) at ipc.c:517
> 517                 rc = qb_ipcc_recv(client->ipc, client->buffer,
> client->buf_size, 500);
> Value returned is $2 = -11
> (gdb) n
> 518                 if(rc > 0 || crm_ipc_connected(client) == FALSE)
> {
> (gdb) p rc
> $3 = -11
> (gdb) n
> 522             } while(time(NULL) < timeout);
> (gdb) n
> 517                 rc = qb_ipcc_recv(client->ipc, client->buffer,
> client->buf_size, 500);
> (gdb) n
> 518                 if(rc > 0 || crm_ipc_connected(client) == FALSE)
> {
> (gdb) p rc
> $4 = -11
> (gdb) n
> 522             } while(time(NULL) < timeout);
> (gdb) n
> 517                 rc = qb_ipcc_recv(client->ipc, client->buffer,
> client->buf_size, 500);
> (gdb) n
> 518                 if(rc > 0 || crm_ipc_connected(client) == FALSE)
> {
> (gdb) p rc
> $5 = -11
> (gdb) n
> 522             } while(time(NULL) < timeout);
> (gdb) n
> 517                 rc = qb_ipcc_recv(client->ipc, client->buffer,
> client->buf_size, 500);
> (gdb) n
> 518                 if(rc > 0 || crm_ipc_connected(client) == FALSE)
> {
> (gdb) p rc
> $6 = -11
> (gdb)
> 
> It seems that lrmd has repeated the reply reception from stonithd
> out of the g_main_loop. Therefore, monitor of Dummy is not performed.
> 
> [root at dev1 ~]# top -bn1
> top - 18:53:31 up 5 days,  8:56,  4 users,  load average: 0.98, 0.63,
> 0.44
> Tasks: 198 total,   2 running, 196 sleeping,   0 stopped,   0 zombie
> Cpu(s):  0.9%us,  1.2%sy,  0.0%ni, 97.9%id,  0.0%wa,  0.0%hi,
>  0.0%si,  0.0%st
> Mem:   5089052k total,  2417444k used,  2671608k free,   266660k
> buffers
> Swap:  1048568k total,        0k used,  1048568k free,  1724616k
> cached
> 
>   PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
> 18589 root      20   0 83932 3448 2548 R 98.2  0.1   4:15.89 lrmd
>     1 root      20   0 19348 1520 1212 S  0.0  0.0   0:00.78 init
>     2 root      20   0     0    0    0 S  0.0  0.0   0:00.00 kthreadd
>     3 root      RT   0     0    0    0 S  0.0  0.0   0:06.93
>     migration/0
>     4 root      20   0     0    0    0 S  0.0  0.0  15:23.59
>     ksoftirqd/0
>     5 root      RT   0     0    0    0 S  0.0  0.0   0:00.10
>     migration/0
> 
> Best Regards,
> Kazunori INOUE
> 
> _______________________________________________
> 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