[Linux-HA] crm crash on centos 4.5

[Andrew Beekhof]

On Feb 7, 2008, at 12:57 AM, Lars Marowsky-Bree wrote:

> On 2008-02-06T15:11:28, Tao Yu <taoyu2007 at gmail.com> wrote:
>
>> Running the heartbeat 2.1.2
>>
>> Core:
>>
>> #0  0x0000003b9602e21d in raise () from /lib64/tls/libc.so.6
>> #1  0x0000003b9602fa1e in abort () from /lib64/tls/libc.so.6
>> #2  0x0000003efe80bc90 in crm_abort () from /usr/lib64/ 
>> libcrmcommon.so.1
>> #3  0x0000003efee01adb in ?? () from /usr/lib64/libccmclient.so.1
>> #4  0x0000003efee021ba in cookie_ref () from /usr/lib64/ 
>> libccmclient.so.1
>> #5  0x0000003b9751ce5f in g_hash_table_size () from /usr/lib64/ 
>> libglib-2.0.so.0
>> #6  0x0000003efee02557 in oc_ev_handle_event () from
>> /usr/lib64/libccmclient.so.1
>> #7  0x000000000040e700 in ?? ()
>> #8  0x0000003efe21776f in g_main_output_is_blocked () from
>> /usr/lib64/libplumb.so.1
>> #9  0x0000003b97526606 in g_main_context_dispatch () from
>> /usr/lib64/libglib-2.0.so.0
>> #10 0x0000003b9752821e in g_main_context_acquire () from
>> /usr/lib64/libglib-2.0.so.0
>> #11 0x0000003b9752858a in g_main_loop_run () from /usr/lib64/ 
>> libglib-2.0.so.0
>> #12 0x0000000000404499 in ?? ()
>> #13 0x0000003b9601c3fb in __libc_start_main () from /lib64/tls/ 
>> libc.so.6
>> #14 0x00000000004040ca in ?? ()
>> #15 0x0000007fbffff8d8 in ?? ()
>> #16 0x000000000000001c in ?? ()
>> #17 0x0000000000000001 in ?? ()
>> #18 0x0000007fbffffb93 in ?? ()
>> #19 0x0000000000000000 in ?? ()
>>
>> ha_debug file isn't created for this date :-( Really do not know why
>> do not have the ha-debug file.
>>
>> Is this a known issue?

Yes.  This is the infamous "CCM counter going backwards" issue.
So it's not actually a crash, its the crmd saying "the world has gone  
crazy and I'm getting out of here right now!" and doing so in a way  
that creates a stack trace for us to examine later.

> Almost impossible to say, you should try to install the
> heartbeat-debuginfo package if CentOS has something like it.
>
> If not, I'd recommend to upgrade to the latest 2.1.3 and hope it is
> fixed ;-)

i believe it was