[ClusterLabs] Possbile cause of long down time of shared IP during manual failover.
lists at alteeve.ca
Tue Jun 2 21:34:39 EDT 2015
On 02/06/15 09:22 PM, Akira Miyazaki wrote:
> Thank you for prompt reply.
> As I'm just user of the enviroment, I can not update softwares by myself.
> I need to ask right pserson to update that.
> To do so, I need explain why need update softwares.
> So I would like to get information if my symptom was fixed already.
> Is it possbile to get such a information somewhere?
Here is a good reason;
Back in 2008, the various people who make High Availability software
decided to merge the two different HA stacks into one. That process
finished when RHEL 7 was released (other distributions merged sooner).
The only distribution still behind is Debian, but the latest release
shipped an incomplete and broken HA stack anyway. So Debian, as it is
now, is not any good for HA at this time.
So now, all current and future development will be working on Corosync
and Pacemaker. These programs reached maturity (in the context of the
new stack) around Pacemaker version 1.1.10 and Corosync v2 (v2.3.4 or
newer being recommended).
Continuing to use old corosync and old pacemaker means you will miss
many bug fixes and getting help from the community will be hard (and get
even harder over time) as fewer and fewer people will be left using it.
This is before even discussing the very large number of bug fixes that
have gone in over the many years since those were old versions released.
If the people who make this decision do not wish to update, even with
this information, then you (or they) will need to spend a fair bit of
time narrowing down the source of a given problem before you can
identify if that problem was fixed later. I think you and they will save
a lot of time moving to the newer stack.
All the best
PS - Use stonith! :)
Papers and Projects: https://alteeve.ca/w/
What if the cure for cancer is trapped in the mind of a person without
access to education?
More information about the Users