[LinuxFailSafe] the cdb doesn't update across the cluster

[Daniel Berg]

Hi,
Thanks for all the help so far.

Finally I got things going. I have defined two nodes and a cluster, I've added the nodes to the cluster and created a HA resourcegroup containing only one IP-adress resource, for testing. I'm having some problems bringing the secondary node online. I run fstask and login to the primary node, I try to start ha_services on the cluster but only one node (the primary) goes online, the secondary gets red and the message "down". In system guard on the sec. node I can see that no ha-processes are starting up as I try to bring it online.
Before I try to start HA the status is "Inactive" and not "Unknown" on the sec. node called "Tyson". The primary node is called "Tore".
Here is the cdbd-log from Tore (prim.), it seems like it can't find the sec. node tyson. : 

linux cdbd  - fs2d_init: cannot open database, error 6
linux cdbd  - Initialization failed
linux cdbd  - fs2d_init: cannot open database, error 6
linux cdbd  - Initialization failed
linux cdbd  - Quorum log last entry at startup is quorum: cluster id: 0x00000000.0x3d1869582c1a7b48, master: 1, sequence: 1, member count: 1, members:  1
linux cdbd  - Added new pool machine Tore (1)
linux cdbd  - Last remote quorum entry for machine Tore (1) is quorum: cluster id: 0x00000000.0x3d1869582c1a7b48, master: 1, sequence: 1, member count: 1, members:  1
linux cdbd  - Added new pool machine Tyson (2)
linux cdbd  - No known remote quorum entry for machine Tyson (2)
linux cdbd  - started (pid 30990)
linux cdbd  - fs2d main thread ready
linux cdbd  - Setting local machine id to 1, old value = 0
Tore cdbd  - Quorum formation waiting for 1 status check
Tore cdbd  - Quorum formation waiting for 1 status check
Tore cdbd  - Pool machine Tyson (2) not responding //This doesn't look good!!!
Tore cdbd  - New quorum requested (not forced)
Tore cdbd  - Need replacement for old quorum: cluster id: 0x00000000.0x3d1869582c1a7b48, master: 1, sequence: 1, member count: 1, members:  1
Tore cdbd  - Proposed new quorum: cluster id: 0x00000000.0x3d1869582c1a7b48, master: 1, sequence: 2, member count: 1, members:  1
Tore cdbd  - New quorum: cluster id: 0x00000000.0x3d1869582c1a7b48, master: 1, sequence: 2, member count: 1, members:  1
Tore cdbd  - New quorum: cluster id: 0x00000000.0x3d1869582c1a7b48, master: 1, sequence: 2, member count: 1, members:  1
Tore cdbd  - Ready and valid new quorum: cluster id: 0x00000000.0x3d1869582c1a7b48, master: 1, sequence: 2, member count: 1, members:  1

Q1:
Do I have to explicitly syncronize the databases across the cluster, and if so how?

Q2:
When defining a node, should the IP-adresses configured on the machine be given, or should i pick some other IP-adresses?
Because when defining an IP-adress resource, the manual says I can't give it the same IP-addr as the ones configured on the machines interfaces.

Q3: 
I have two ethernet-interfaces on the machine, and I would like for one to a backup in case the first one goes down. Should the secondary interface be "down" in the initial state, and if so, doesn't this trigger a failover instantly?

Thank you very much//Daniel

-------------------------------------------
Daniel Berg
CC-systems AB
Väktargatan 44B nb
754 22 Uppsala
tel: 018-25 75 30
mob: 070-634 5556


-- 
__________________________________________________________
Sign-up for your own FREE Personalized E-mail at Mail.com
http://www.mail.com/?sr=signup

Save up to $160 by signing up for NetZero Platinum Internet service.
http://www.netzero.net/?refcd=N2P0602NEP8