Re: How can I load ha plugin?

"jin li" <[email protected]> wrote on 03/08/2007 01:16:07 AM:

> hi, Steve
> thank you very much for your reply
> I use heartbeat-1.2.4
> and ,i delete the  /lib/evms/2.5.5/.kdbgrc.ha-1.1.1.so ,so there is
> no more the first error
> and ,this error still there:
> Engine: The plug-in Linux-HA in module /lib/evms/2.5.5/ha-1.1.1.so
> failed to load.
> The plug-in's setup_evms_plugin() function failed with error code
> 19: No such device.
>
> and I used the debug,and when I check the var/log/evms-engine.log,I
> found something :
>  3月 08 13:38:12 localhost.localdomain _8_ Engine: status_message:
> Setup plug-in: Linux-HA Cluster Manager
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: ece_init: Enter.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: global_init: Enter.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: cleanup_membership:
Enter.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: cleanup_memb_crc:
Enter.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: cleanup_memb_crc:
Exit.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: cleanup_membership:
Exit.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: cleanup_messaging:
Enter.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: cleanup_messaging:
Exit.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: cleanup_eventq:
Enter.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: cleanup_eventq:
Exit.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: unregister_all:
Enter.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: unregister_all:
Exit.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: source_delete_all:
Enter.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: source_delete_all:
Exit.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: mainloop_cleanup:
Enter.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: mainloop_cleanup:
Exit.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: frag_cleanup: Enter.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: frag_cleanup: Exit.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: condition_init:
Enter.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: condition_init:
Exit.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: global_init: Exit.
>  3月 08 13:38:12 localhost.localdomain _7_ Linux-HA: sm_cnct: Enter.
>  3月 08 13:38:13 localhost.localdomain _1_ Linux-HA: sm_cnct: Error:
> Daemon may not be running
>  3月 08 13:38:13 localhost.localdomain _7_ Linux-HA: sm_cnct: Exit.
> Return pointer = (nil)
>  3月 08 13:38:13 localhost.localdomain _1_ Linux-HA: ece_init: Error
> connecting to ECE Daemon
>  3月 08 13:38:13 localhost.localdomain _7_ Linux-HA: ece_init: Exit.
> Return value = 19
>  3月 08 13:38:13 localhost.localdomain _7_ Engine: engine_user_message:
Enter.
>  3月 08 13:38:13 localhost.localdomain _0_ Engine:
> engine_user_message: Message is: Engine: The plug-in Linux-HA in
> module /lib/evms/2.5.5/ha- 1.1.1.so failed to load.  The plug-in's
> setup_evms_plugin() function failed with error code 19: no such device
>  3月 08 13:38:13 localhost.localdomain _7_ Engine:
> engine_user_message: Exit.  Return value is 0.
>  3月 08 13:38:13 localhost.localdomain _7_ Engine: release_plugin: Enter.
>  3月 08 13:38:13 localhost.localdomain _7_ Engine: unload_module: Enter.
>  3月 08 13:38:13 localhost.localdomain _8_ Engine: unload_module:
> Unload module /lib/evms/2.5.5/ha-1.1.1.so.
>  3月 08 13:38:13 localhost.localdomain _8_ Engine: unload_module:
> Issuing dlclose() on handle 0x80e6aa8.
>  3月 08 13:38:13 localhost.localdomain _7_ Engine: unload_module: Exit.
>  3月 08 13:38:13 localhost.localdomain _7_ Engine: release_plugin:
> Exit.  Return value is 0.
> I do not understand what the meaning of the line:Daemon may not be
> running.What is the daemon?
> Is it heartbeat? I can use this:/etc/init.d/heartbeat start to start
> it, and by use this:ps aux |grep heartbeat  I found the heartbeat running
well
> Or is it evmsd? When I use evmsd , I got this: Deamon:The plug-in
> Linux-HA in module /lib/evms/2.5.5/ha-1.1.1.so failed to load.
> The plug-in's setup_evms_plugin() function failed with error code
> 19: No such device. Deamon:There is no cluster manager plug-in
> loaded in this system.
>
> So, I need more help
>
> Hoping reply !
> Thanks again！
>
>
>
>
> 2007/3/8, Steve Dobbelstein <[email protected]>:
> "jin li" <[email protected]> wrote on 03/06/2007 09:52:58 PM:
>
> > Hi everbody,
>
> > I want to configure EVMS for a cluster, then I followed all the
> > steps in http://evms.sourceforge.net/install/cluster.html
> > But after that, when I entered EVMS interface by evsmgui command,
> there was a
> > alert message:
> > Engine: Error loading /lib/evms/2.5.5/.kdbgrc.ha- 1.1.1.so:
> > /lib/evms/2.5.5/.kdbgrc.ha- 1.1.1.so: invalid ELF header
> > Engine: The plug-in Linux-HA in module /lib/evms/2.5.5/ha- 1.1.1.so
> > failed to load.
> > The plug-in's setup_evms_plugin() function failed with error code
> > 19: No such device.
> > I have adder these lines into ha.cf:
> > respawn hacluster /usr/lib/heartbeat/ccm
> > respawn root /sbin/evmsd
> > apiauth evms uid=root
> > apiauth ccm uid=hacluster
> > But the alert still exist
> > I don't know why, I have started heartbeat by /etc/init.
> d/heartbeat restart.
> > I am using heartbeat-1.2.4, evms-2.5.5, and my os is fedora core 4
> > kernel is 2.6.12.
> > I also found the option Setting -> Node Administrated... was disable
> >
> > Anyone can tell me how can I resolve the problem?
> > Hoping reply,Thanks first!

> My guess is that the two error messages are two separate errors.
>
> The "Error loading /lib/evms/2.5.5/.kdbgrc.ha- 1.1.1.so" means,
> naturally, that EVMS could not load the file /lib/evms/2.5.5/.kdbgrc.ha-
> 1.1.1.so as a plug-in.   I have never heard of a file by that name.
> I suspect it is junk in the /lib/evms/ directory.  You can simply
> delete the file to avoid the error message.  This error should not
> cause any subsequent errors.  The Engine failed to load a bad file
> and then continued.
>
> The HA plug-in will return error code 19 when it fails to connect to
> heartbeat.  The connection can fail for a variety of reasons.  Try
> running with the debug level set to "debug", for example, "evmsgui -
> d debug".  Then look at the log file (default is /var/log/evms-
> engine.log) and see if there are any debug messages that say why the
> HA plug-in couldn't connect to heartbeat.
>
> What version of Linux HA are you using?  If you are using version 2
> or later, try using the hb2 plug-in instead.  You will have to
> configure EVMS to build the hb2 plug-in instead of the ha plug-in.
> run "./configure --disable-ha --enable-hb2 <your other config
> options>".  Then rebuild and reinstall EVMS.
>
> Hope this helps.
>
> Steve D.

Hello again.

Sorry, I should have thought about the daemon.  I have been away from the
HA code for a while.

EVMS has a daemon, evmsd, which is the thing that actually connects to
heartbeat.  When you bring up an EVMS user interface (evms, evmsn, evmsgui)
the Engine connects to the EVMS daemon, not to heartbeat.

One of the purposes of the daemon is to assist in coordinating
configuration changes across all the nodes in the cluster.  The daemon
should be running on each node in the cluster.  I see that you have
heartbeat configured to launch the EVMS daemon, /sbin/evmsd.  That's good.

Here are a few things you can check.

Is the daemon still running?  You configured heartbeat to respawn evmsd, so
it should be running.  But if I recall correctly, if evmsd continues to
fail to run, heartbeat will eventually give up on trying to respawn it.
Check to see if evmsd is running.

Another thing to check is the log file for the daemon.  The default is
/var/log/evms-daemon.log.  If evmsd fails to run there may be some messages
in the log file.  Then again there may not be any messages.  The logging
level for evmsd is kept to a minimum since it is always running.  You can
set the log level to "debug" for evmsd.  Run "evmsd -d debug".  You may
have to change your heartbeat configuration to add the debug level option.
Once you are finished with the debugging remember to set the logging level
back to "default".  Since evmsd is always running, it will keep writing
debug messages to the log file and consume more and more disk space.

If evmsd is failing to load try running "evmsccm -l".  evmsccm connects to
heartbeat and queries the CCM membership.  It connects to heartbeat in the
same way evmsd does.  If it fails to connect it should display a message
saying what went wrong.

Hope this helps.

Steve D.

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

_______________________________________________
Evms-devel mailing list
[email protected]
To subscribe/unsubscribe, please visit:
https://lists.sourceforge.net/lists/listinfo/evms-devel