On 4/12/07, 池田淳子 [EMAIL PROTECTED] wrote:
Hi all,
I'm newbie, and trying to understand how or when Split-Brain happens.
when 1 or more nodes can't communicate with each other
there is no connection between this and an operation's start-delay
what you *might* be seeing is an old bug that
池田淳子 wrote:
Hi all,
I'm newbie, and trying to understand how or when Split-Brain happens.
As trial, I run Dummy resource for now.
Split-brain occurs when there is a total communication failure (from the
heartbeat perspective) between at least two different cluster nodes.
Heartbeat version
Hi,
Just like to know your opinion about the following. A pure Linux shop
would of course definitely go for Heartbeat as the solution for high
availability. However, in an environment that comes from Novell's
NetWare, Novell Cluster Services (NCS) would be the best choice,
especially if running
On 4/11/07, Lars Marowsky-Bree [EMAIL PROTECTED] wrote:
On 2007-04-10T18:29:10, Andrew Beekhof [EMAIL PROTECTED] wrote:
Apr 10 17:30:25 ha-test-1 process[26425]: Returnig 7
Apr 10 17:30:40 ha-test-1 process[26493]: Maintainance =
Apr 10 17:30:40 ha-test-1 process[26493]: OCF_RESKEY_probe = 1
i hate to pester, but where are the fail counts kept track of and what
maintains
them?
they are stored in the status section and are maintained by the
tengine process (which increases it whenever a monitor action fails)
there is also a CLI tool called crm_failcount that can be used to view
NCS has better integration with EVMS, and has data-network heartbeat. It
does not therefore require STONITH.
It has had much more testing than HB for large clusters as well. 20+
node clusters are not uncommon.
Yan
Sander van Vugt wrote:
Hi,
Just like to know your opinion about the
Oh - and I forgot... UI was _much_ nicer last time I looked. And nowhere
near as buggy as the HB GUI.
Yan Fitterer wrote:
NCS has better integration with EVMS, and has data-network heartbeat. It
does not therefore require STONITH.
It has had much more testing than HB for large clusters as
Andrew Beekhof wrote:
On 4/11/07, Terry L. Inzauro [EMAIL PROTECTED] wrote:
list,
this is a continuation of another thread that was started a few weeks
back. the original thread was
started in regards
to the setup of pingd. this thread is in regards to pingd not being
able to start for
Andrew Beekhof wrote:
On 4/11/07, Terry L. Inzauro [EMAIL PROTECTED] wrote:
list,
this is a continuation of another thread that was started a few weeks
back. the original thread was
started in regards
to the setup of pingd. this thread is in regards to pingd not being
able to start for
Terry L. Inzauro wrote:
Andrew Beekhof wrote:
On 4/11/07, Terry L. Inzauro [EMAIL PROTECTED] wrote:
list,
this is a continuation of another thread that was started a few weeks
back. the original thread was
started in regards
to the setup of pingd. this thread is in regards to pingd not
Sander van Vugt wrote:
Hi,
Just like to know your opinion about the following. A pure Linux shop
would of course definitely go for Heartbeat as the solution for high
availability. However, in an environment that comes from Novell's
NetWare, Novell Cluster Services (NCS) would be the best
Bernd Eichenberg wrote:
Hi at all,
I've to warn you, because I'm a newby at HA and I'm german
with very small english skillz.
Hope you understand my problem and thank for that.
My Problem is, there is no failback after the 1.node is valid
again.
1. node is Suse 8 with heartbeat that
Dominik Klein wrote:
Hi
I use heartbeat 2.0.7 from openSuSE 10.2
10.250.250.27 is master
10.250.250.28 is backup
They are connected with direct ethernet cable
When watching UDP heartbeats with tshark, my master machine says this
(checksum errors):
16:27:24.140509 10.250.250.28 -
On 2007-04-12T08:58:15, Andrew Beekhof [EMAIL PROTECTED] wrote:
So I thought that probe is maybe never unset
correct
http://www.osdl.org/developer_bugzilla/show_bug.cgi?id=1479
lmb - not the same thing. the resource is not deleted in between the
two types of monitor calls.
this
Hi All,
We are using linux ha for achieving HA solution for a 2 node system. We have
observed memory leaks in the crmd process and seen it grow to beyond 700MB.
I am currently running 2.0.8 with a patch that fixes a few memory leak
issues in crmd. I am trying to use valgrind and efence to nail
15 matches
Mail list logo