[Linux-HA] [right one] Please can anybody help: switch from v1 to v2 with 25 nodes

[Andrew Beekhof]

On 3/9/07, Aleksandar Lazic <al-linuxha at none.at> wrote:
> Dear HA Community,
>
> please apologize my unhappy mail stile but I'am sitting here very
> hopeless in Saudi-Arabia and try to configure the cluster described
> later in this mail.
>
> The first mail have wrong from mail-addresse, sorry :-(
>
> I have setup some v1 cluster with 2 nodes and IP and application take
> over.
> Now I have try to make  a cluster with 25 nodes.
> I have added all the nodes into the ha.cf file.
> I have added crm yes into the same file.
> With the haclient.py I see all of the nodes.
> But there are some errors which I can't handle by my own.

can i suggest you start with something slightly smaller?
try a two or three node "crm" cluster with an ip address and work up from there.

as for the xml below... did you make that up?

try looking here:
   http://www.linux-ha.org/v2/dtd1.0/annotated
for a description of the actual syntax.

there is also a haresource2cib.py script which will take a v1 config
and create a v2 equivalent.

> The 64bit environment is:
> HP-Blades
> Redhat Linux ES 4.4
> heartbeat 2.0.7 from a centos mirror
>
> 22 Active Server which are connected to an HP-SAN.
> Every of the server have a LVM for /var/lib/mysql and a /data/Data1.
>
> The LVM names are:
> $HOSTNAME_VGDB/$HOSTNAME_LVDB
> $HOSTNAME_VGDAT/$HOSTNAME_LVDAT
>
> The 3 Standby server are able to import the VGs and LVs, after a clean
> umount and vgexport.
> Know the single steps works for them self.
>
> I have try to make the setup with the haclient.py but I get always the
> "qourum"-error (to write you need a qourum). I have seen in the
> ML-Archive that I should not change the have_quorum to true.
> I have also seen this: http://www.linux-ha.org/QuorumServerGuide it
> looks to me that this is only possible with the 2.0.8 version.
>
> How can I add a qourum in 2.0.7 without recompiling?
> I have try to create a rpm from 2.0.8 src with the tsa_plugin but is was
> unable to do this because of 2 reasons:
>
> 1.) the ha_mgmt_client.i was not in the tgz.
>   I get it from
>   http://cvs.linux-ha.org/viewcvs/viewcvs.cgi/linux-ha/tsa_plugin/
>
> 2.) I was unable to create linuxhafla.jar due the fact that I don't have
>     find where I can get this java package "com.ibm.eez.sdk" or the
>     linuxhafla.jar. I have get the javafiles from the cvs.
>
> Please help me to solve this problem :-(
>
> What is the target:
>
> 22 aktive server with 1 vip per server, 2 lvms connected to HP-SAN, 1
> /etc/init.d/SCRIPT.
> The application is on all server the same.
>
> there are 3 standby servers which should takeover the vip, lvms and
> start the /etc/init.d/SCRIPT. Maybe that the lvm takeover be handled
> with the SCRIPT.
>
> I have try with the haclient to add a group to the group the resouce
> IPaddr and add a "place" on which this "resouce" should be.
>
> What I need is like a template for the cib.xml due the fact that I think
> I'am in the right direction but can't find *the* way ;-), please correct
> me if I'am wrong.
>
> I know that this is a very *big* request and I don't want to get your
> commiserate or steal your time to describe the hopless situation here or
> how this was happend.
>
> I hope that anybody take pity to me.
>
> What I think that the cib.xml should works like this but I don't know
> the right syntax :-(
>
> -----------------------------------------------
> <server id="01" name"Server01" default="active" status="running">
>  <group_TO_handle_01/>
> <server/>
> <server id="02" name"Server02" default="active" status="running">
>  <group_TO_handle_02/>
> <server/>
> .
> .
> .
> <server id="22" name"Server22" default="active" status="running">
>  <group_TO_handle_22/>
> <server/>
>
> <server id="23" name"Server23" default="standby" status="running">
>  <group_TO_handle_01/>
>  <group_TO_handle_02/>
>     .
>     .
>  <group_TO_handle_22/>
> <server/>
>
> <server id="24" name"Server24" default="standby" status="running">
>  <group_TO_handle_01/>
>  <group_TO_handle_02/>
>     .
>     .
>  <group_TO_handle_22/>
> <server/>
>
> <server id="24" name"Server24" default="standby" status="running">
>  <group_TO_handle_01/>
>  <group_TO_handle_02/>
>     .
>     .
>  <group_TO_handle_22/>
> <server/>
>
> <group_TO_handle_01 id="01" default="active" on_server="Server01">
>   <IPaddress/>
>   <LVM_IF_NOT_HANDLED_BY_SCRIPT/>
>   <SCRIPT/>
> <group_TO_handle_01/>
> .
> .
> .
> <group_TO_handle_22 id="22" default="active" on_server="Server22">
>   <IPaddress/>
>   <LVM_IF_NOT_HANDLED_BY_SCRIPT/>
>   <SCRIPT/>
> <group_TO_handle_22/>
> -----------------------------------------------
>
> Sorry the mail is longer then I thought.
>
> BR
>
> Aleks
> _______________________________________________
> Linux-HA mailing list
> Linux-HA at lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
>