I have a two node webserver-cluster running GPFS for
shared filesystem, and heartbeat for active/active ip-address
failover. Sometimes it can take a few moments before the shared
filesystem is mounted, so it would be good if heartbeat could
wait until the filesystem was mounted before failing back
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Fri, May 9, 2008 at 12:30 AM, Dominik Klein wrote:
> Do you need a replication slave?
>
> Otherwise, I'd personally rather go with drbd instead of mysql replication.
> It is synchronous in constrast to the asynchronous mysql replication and
> drbd
Hi again,
Thanks for your comments.
However, I'm afraid I did try that.
in the URL you posted, the steps are pretty much detailed, but more
specifically for versions > 2.0.8...
In any case, I tried the step by step about three times, re-ordering the
steps after it failed for the first time, witho
On Fri, May 9, 2008 at 12:47 PM, Nikola Ciprich <[EMAIL PROTECTED]> wrote:
> Hello Junko,
> thanks for the answer. I gave pacemaker a try, and I think I'll stick to it.
> But unfortunately it didn't solved my problem, some resources are still
> being restarted for some reason.
can you use hb_repor
Dear Dejan (and all :))
I will try again (taking a closer look at the pingd question) to apply the
FAQ steps on replacing a node...
In any case, I won't be able to do that before monday 19 (not next
monday...) so I will come back to you guys with an update by then...
Many thanks, again, for the a
Hello Junko,
thanks for the answer. I gave pacemaker a try, and I think I'll stick to it.
But unfortunately it didn't solved my problem, some resources are still
being restarted for some reason.
ie if I put one node (vbox4) into standby mode, clvmd gets restarted on the
other.
log output:
May
Hi,
On Fri, May 09, 2008 at 11:46:29AM +0200, Stallmann, Andreas wrote:
> Hi there,
>
> were still in deep sh** with heartbeat and drbd in a split brain
> szenario.
>
> We have the following set up:
>
> - A two node active/passive cluster (heartbeat 2.1.3 without crm)
> - Dopd with drbd-peer-ou
Hi,
On Fri, May 09, 2008 at 12:24:39PM +0200, Stallmann, Andreas wrote:
> Hi!
>
> An other issue adding to the problem described before:
>
> For testing purposes, we set in ha.cf:
>
> auto_failback off
>
> Still, when our old primary comes back, it takes over the ressources!
> Gnagnagnagna...!
Hi!
An other issue adding to the problem described before:
For testing purposes, we set in ha.cf:
auto_failback off
Still, when our old primary comes back, it takes over the ressources!
Gnagnagnagna...!
Please, make it stop!
*sigh*
It seems that the setting has no consequence at all!
Thanks
Hi there,
were still in deep sh** with heartbeat and drbd in a split brain
szenario.
We have the following set up:
- A two node active/passive cluster (heartbeat 2.1.3 without crm)
- Dopd with drbd-peer-outdater (the newest ones, patched).
- Ipfail
Still, if we disconnect one host from the netw
Hi,
On Fri, May 09, 2008 at 11:23:49AM +0200, Niko wrote:
> Dear all,
>
> I have the following environment set up:
>
> ===
> OS: RHEL 4 update5
> HeartBeat: 2.0.8
> Node type: Iptables + FwBuilder
> ===
> Problem origin:
> The environment w
Hi,
On Thu, May 08, 2008 at 10:33:50AM -0500, Matt Zagrabelny wrote:
> Greetings,
>
> I have written a STONITH resource agent and am seeking clarification on
> the following parameter attributes:
>
> required and default
>
> Specifically, I would like to be ensured that there are some variables
Dear all,
I have the following environment set up:
===
OS: RHEL 4 update5
HeartBeat: 2.0.8
Node type: Iptables + FwBuilder
===
Problem origin:
The environment was working fine before an incorrect shut down of one of the
machines (for balance
Hi there
Strange, I've been using this feature for several months...
Is your Heartbeat version too old?
Regards,
Chun Tian (binghe)
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
This is for v2 only, against the ldirectord OCF script versioned:
$VERSION_STR = "Linux Director v1.186-ha-2.1.2
Hi Aravindo
Aravindo wrote:
Dear List readers
I have a MasterSlave configuration with drbd, similar to
http://www.linux-ha.org/DRBD/HowTov2. That works great.
As it is a Master-Slave configuration, there is one instance on both nodes.
crm_mon looks like this:
Node: node1 (aa86a87b-355a-4868-
Dear List readers
I have a MasterSlave configuration with drbd, similar to
http://www.linux-ha.org/DRBD/HowTov2. That works great.
As it is a Master-Slave configuration, there is one instance on both nodes.
crm_mon looks like this:
Node: node1 (aa86a87b-355a-4868-8a99-e7fdc024cf91): online
16 matches
Mail list logo