Forum
Welcome, Guest
Username: Password: Remember me
This is the optional category header for the Suggestion Box.
  • Page:
  • 1
  • 2

TOPIC:

Fujitsu iRMC fencing scripts 7 years 10 months ago #835

In the HA--Lizard_1.8.7 doc I could read

FENCE_MIN_HOSTS
Do not allow fencing when fewer than this number of hosts are remaining in the pool. Default value = 2 which is optimized for a 2-node pool.


Currently I use the default FENCE_METHOD=POOL

I could use FENCE_METHOD=IRMC with my two node pool no-san setup.

My question is, should I better change to IRMC instead of POOL and should I set FENCE_MIN_HOSTS=1 in this case?

BR Andreas

Please Log in or Create an account to join the conversation.

Fujitsu iRMC fencing scripts 7 years 10 months ago #836

  • Salvatore Costantino
  • Salvatore Costantino's Avatar
  • Offline
  • Posts: 722
You should leave min_hosts set to 2. The second vote will come from checking heuristic_ips to ensure that the surviving host is really live on its management network.

Please Log in or Create an account to join the conversation.

Fujitsu iRMC fencing scripts 7 years 10 months ago #838

Which fencing method would you prefer to use? (If you would operate Fujitsu nodes with iRMC.)

I am asking this myself as it was mentioned in another thread about a split brain szenario that HW STONITH must be used to prevent that the slave starts a vm before the master is really off.

Is the IRMC logic an advantage compared to the POOL fence method.

If someone would share some thoughts about it...

Please Log in or Create an account to join the conversation.

Last edit: by ajmind.
  • Page:
  • 1
  • 2