On 2008-08-28T19:50:42, Andrew Beekhof [EMAIL PROTECTED] wrote:
that sounds great. Do you know how much this packages differ from the
suse-10.2 packages???
different dependancies. not sure i'd mix them personally
Not for anything worth running in production, for sure not.
I'd honestly
On 2008-08-26T20:10:24, Todd, Conor [EMAIL PROTECTED] wrote:
I'm about to switch the primary interface over to static IPs and put the
hostname to IP mapping in /etc/hosts so that the cluster nodes will always be
able to resolve their hostnames.
Static IPs? Were the hostnames periodically
On 2008-08-26T11:52:35, Matt Zagrabelny [EMAIL PROTECTED] wrote:
Set use_logd yes in the ha.cf file and log to syslog; the logs are so
much more readable.
Out of curiosity, why do you say that using syslog makes the logs so
much more readable?
Well, of course, the line format is a personal
On 2008-08-27T16:49:23, Mega Mailingliste [EMAIL PROTECTED] wrote:
why do you want these specific versions?
Because this is one of these versions which runs stable on our environment.
The other versions before, got us into trouble with some failovers and other
stuff. So why we should use a
On 2008-08-27T11:36:37, Enrique José Hernández Blasco [EMAIL PROTECTED] wrote:
Write this into the accoring OCF RA scripts. It is plain bash, so
shoild not be too difficult.
I've set up a directory per start/stop hook and OCF RA and run-parts
on there but I think it could be cooler o:) to
On 2008-08-26T14:00:35, Ciro Iriarte [EMAIL PROTECTED] wrote:
I was talking about the combination of OCFS2 with HA for membership,
for other tasks I think it'll be as useful as today...
We will continue to support the integration between OCFS2 and the
user-space cluster membership layer in
On 2008-08-27T18:42:41, Mega Mailingliste [EMAIL PROTECTED] wrote:
Right that's true, but where're using the suse build service since 3 or 4
installations and, I know thats my mistake, we forgot to save the rpm's
anywhere. How should know thats the rpm's changing so fast.
Sorry, the build
On 2008-08-27T18:48:01, Andreas Mock [EMAIL PROTECTED] wrote:
Despite that you have to emphasize the following:
What do you do if the piece of code in the
do-it-after-resource-start/stop-hock
has an error? What do you send back as result code? What happens if the
On 2008-08-25T23:50:26, Todd, Conor [EMAIL PROTECTED] wrote:
My understanding with heartbeat is that the bcast directive tells it
to use UDP broadcast on the interfaces in the order in which they're
given. UDP broadcast doesn't rely on DNS, so I'm wondering why a loss
of DNS resolution would
On 2008-08-26T13:34:31, Brian Reichert [EMAIL PROTECTED] wrote:
Are you using IP addresses to describe the other nodes, or are you
using hostnames?
The docs say to use IP addresses, but I've seen hostnames work.
But these do not prevent nodes from seeing each other. The media
channels do
On 2008-08-25T11:29:58, Adrian Chapela [EMAIL PROTECTED] wrote:
And if I run:
./ConfigureMe make --enable-libc-malloc
the output error is the same.
Have you another idea ?
Yes. Build with --enable-fatal-warnings=no
___
Linux-HA-Dev:
On 2008-08-25T16:43:14, Adrian Chapela [EMAIL PROTECTED] wrote:
Have you another idea ?
Yes. Build with --enable-fatal-warnings=no
I was sure about I tried this option but I try again now and it was
compiling OK.
The problem was caused by an incompatible type definition - cl_malloc.h
On 2008-08-25T18:05:49, Adrian Chapela [EMAIL PROTECTED] wrote:
The problem was caused by an incompatible type definition - cl_malloc.h
used unsigned long instead of size_t.
Since cl_malloc is now gone from dev, the next release should do without
this warning.
Thank you!
Actually, I spoke
On 2008-08-25T14:04:14, Christoph Eßer [EMAIL PROTECTED] wrote:
Hi there,
meanwhile I managed getting a logfile by a simple reboot of both nodes.
The real problem I have is that Heartbeat doesn't start a configured
Filesystem resource. The logfile entry is rather cryptic and doesn't
help
On 2008-08-25T16:29:14, Klaus Jagemann [EMAIL PROTECTED] wrote:
heartbeat[1539]: 2008/08/25_15:09:12 ERROR: Cannot write to media pipe 2:
Resource temporarily unavailable
heartbeat[1539]: 2008/08/25_15:09:12 ERROR: Shutting down.
Due to the size of the logfile (1.2MB), I will only attach
On 2008-08-25T08:38:19, Chris Conley [EMAIL PROTECTED] wrote:
Sorry if I wasn't clear enough. EC2 is the cloud computing service provided
by Amazon: http://www.amazon.com/gp/browse.html?node=201590011
Basically, I have heartbeat running on two virtual machines in the EC2
cloud. If one of
On 2008-08-24T07:50:42, Todd, Conor [EMAIL PROTECTED] wrote:
I'm wondering if I upgrade one of my cluster nodes to 2.1.4, if the
rest of the nodes will still accept it as part of their cluster? I'd
rather not have to start from scratch in order to upgrade.
In theory, that should work.
On 2008-08-22T11:18:07, Paul Walsh [EMAIL PROTECTED] wrote:
pils.c:244: error: initialization from incompatible pointer type
pils.c:245: error: initialization from incompatible pointer type
make[2]: *** [pils.lo] Error 1
make[2]: se sale del directorio
On 2008-08-22T09:20:02, Ronny Becker [EMAIL PROTECTED] wrote:
Hello,
I am trying to configure a fencing after failover and cannot find a way to
do this.
I am using heartbeat 2.1.4.
This is what I have configured:
- resource_stickiness is about 99
- rosource_failure_stickiness is -50
-
On 2008-08-22T19:40:52, Mike Sweetser - Adhost [EMAIL PROTECTED] wrote:
I want to do the following:
* Remove no_r1_on_gamma
* Remove run_r1
* Create a new constraint called run_r1 with a score of 100 for gamma
* Create a new constraint called no_r1_on_alpha with a score of
-INFINITY for
On 2008-08-22T17:20:40, Kevin Harms [EMAIL PROTECTED] wrote:
I'm wondering if anyone has any experience with the following case. The
STONITH utility uses IPMI to reboot machines and works via a username and
password. This username and password are in the cib.xml. Beyond the standard
Dear readers,
this year's Linux Kongress, held from 7th to 10th October in Hamburg,
features 3 tutorials which explain and use Linux-HA as their clustering
technology.
Congratulations and thanks to Ralph Dehner, Lars Ellenberg, Joerg
Jungermann, Maximilian Wilhelm. It is wonderful to see the
On 2008-08-22T13:53:08, Adrian Chapela [EMAIL PROTECTED] wrote:
pils.c -fPIC -DPIC -o .libs/pils.o
cc1: warnings being treated as errors
pils.c:244: error: initialization from incompatible pointer type
pils.c:245: error: initialization from incompatible pointer type
make[2]: *** [pils.lo]
On 2008-08-22T15:09:01, Adrian Chapela [EMAIL PROTECTED] wrote:
Lars Marowsky-Bree escribió:
On 2008-08-22T13:53:08, Adrian Chapela [EMAIL PROTECTED] wrote:
pils.c -fPIC -DPIC -o .libs/pils.o
cc1: warnings being treated as errors
pils.c:244: error: initialization from incompatible
On 2008-08-22T16:35:45, Ben Beuchler [EMAIL PROTECTED] wrote:
I apologize for the cross-post if that's not kosher. It appears
linux-ha still gets considerably more traffic than the pacemaker list.
Are these functionally equivalent on a cluster with symmetric-cluster=false?
This is the
Hi,
in preparation for 3.0.0, which is bound to happen eventually when it is
ready, I am happy to announce 2.99.0 as a _beta_ version.
2.99.0 introduces significant changes:
- 2.99.0 removes code now maintained outside heartbeat such as the CRM
(now Pacemaker), thus everything which depends
On 2008-08-19T15:02:09, Rasto Levrinc [EMAIL PROTECTED] wrote:
- Remove code which depends on pacemaker (mgmt, dopd, CIM/SNMP)
- Remove unmaintained code (telecom/).
dopd does not depend on crm directly, but it uses clplumbing. Can it stay
there?
Ah, right, dopd stays then I think.
Hi,
where should we maintain CTS?
Pacemaker has it's own fork of it, at this point in time.
But I think CTS makes sense for heartbeat to have too, as _someone_
might still care for the v1 functionality, and it might make sense to
test just the cluster layer w/o resources.
My preferred approach
On 2008-08-21T20:43:55, Andrew Beekhof [EMAIL PROTECTED] wrote:
My preferred approach would be to create heartbeat-cts as a package, and
put CTS in there; and then, Pacemaker would just drop in the overlays
it needs.
wont work as you'd have multiple packages with the same files.
On 2008-08-20T18:33:47, Robert [EMAIL PROTECTED] wrote:
Well, from my perspective - which of course is the perspective of a user
and not the developer of heartbeat - it works. By works I mean, when I
pull the cable between the datacenters, the resources are active on one
node only (the
On 2008-08-20T23:35:36, Robert Heinzmann (ml) [EMAIL PROTECTED] wrote:
I'm getting confused even more while going through some postings and google
findings regarding Linux split site setups and quorum server. Alan writes a
lot about quorum server on his blog:
That's something you'd have to
On 2008-08-20T17:33:01, Ben Beuchler [EMAIL PROTECTED] wrote:
I'm attempting to build debs of this version of Heartbeat on Ubuntu 8.0.4:
http://hg.linux-ha.org/dev/archive/tip.tar.bz2
$ cd Linux-HA-Dev-b32ca6086e32
$ dpkg-buildpackage -rfakeroot
...
All goes well until it tries to
On 2008-08-21T13:38:14, Robert [EMAIL PROTECTED] wrote:
So if quorumd works well in certain environmnts can it be used ? (e.g. non
STONITH configurations).
Sure. 2 Nodes without STONITH, and unable to recover from stop failures
- now that makes for a stable, reliable, and supportable
On 2008-08-21T13:47:07, Andreas Kurz [EMAIL PROTECTED] wrote:
Have you considered to install an additional cluster node as a Quorumserver
which runs no services ?
Doesn't help the STONITH issue at all. Doesn't work in split-site
configurations; the site with N/2+1 nodes will always win, the
On 2008-08-21T13:56:48, Robert [EMAIL PROTECTED] wrote:
Another way could also be, to just setup a 2 node cluster with stonith via
meatware. This allows server maintainance, but automatic failovers have to
be administrator controlled - that's acceptable for the disaster case I
guess.
Hi,
in preparation for 3.0.0, which is bound to happen eventually when it is
ready, I am happy to announce 2.99.0 as a _beta_ version.
2.99.0 introduces significant changes:
- 2.99.0 removes code now maintained outside heartbeat such as the CRM
(now Pacemaker), thus everything which depends
On 2008-08-21T14:52:41, Andreas Kurz [EMAIL PROTECTED] wrote:
Yes, that should work.
But so would already simple majority quorum; set ignore_quorum=freeze
^
by default, and when you want one site/node to continue, force it to
ignore.
Hmm ...
On 2008-08-21T14:25:36, Robert [EMAIL PROTECTED] wrote:
And what about three sides ?
NodeA - DC-A
NodeB - DC-B
NodeQuorum - DC-C
?
Fencing still is not addressed.
Regards,
Lars
--
Teamlead Kernel, SuSE Labs, Research and Development
SUSE LINUX Products GmbH, GF: Markus Rex, HRB
On 2008-08-21T11:44:42, [EMAIL PROTECTED] wrote:
Sorry for my first mail with an incorrect adress... :(
So this post is the second
Hi!
I've set up an small Linux-HA Projekt for myself.
I have two Servers which should provide one IP
Server1: 10.0.0.4
Server2: 10.0.0.5
On 2008-08-21T15:14:08, Michael Schwartzkopff [EMAIL PROTECTED] wrote:
Am Donnerstag, 21. August 2008 14:50 schrieb Lars Marowsky-Bree:
Hi,
in preparation for 3.0.0, which is bound to happen eventually when it is
ready, I am happy to announce 2.99.0 as a _beta_ version.
(...)
Hi,
I
On 2008-08-20T16:33:44, Simon Horman [EMAIL PROTECTED] wrote:
I realise that the whole point of 3.0.0 is to produce
a lha release that depends on pacemaker.
I think you meant the right thing, but the phrasing would be more like
a lha release which pacemaker can depend on ;-)
When I read your
On 2008-08-20T11:22:25, Wolfram Schlich [EMAIL PROTECTED] wrote:
* Robert [EMAIL PROTECTED] [2008-08-19 15:38]:
When thinking about splitting the packages, I think a seperate RA package
is useful. Most changes are done at the RA level (new versions of supported
applications, bug fixes
On 2008-08-20T13:18:42, Wolfram Schlich [EMAIL PROTECTED] wrote:
But yes, clearly we hope to have more independently packaged resource
agents one day; it would be so cool if they were instead shipped with
the service package (and not our problem to maintain! ;-). That's a bit
beyond of
On 2008-08-20T11:18:12, Junko IKEDA [EMAIL PROTECTED] wrote:
Is it hard to replaced the %run_ldconfig macro with /sbin/ldconfig for
RedHat?
No, it's fixed.
___
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org
On 2008-08-20T14:17:17, Wolfram Schlich [EMAIL PROTECTED] wrote:
How the service provider arranges their packaging and upgrade strategy
is their concern, not ours ;-) How often do you upgrade init scripts
without the service?
An OCF RA is not just an init script -- please do not forget the
On 2008-08-20T18:05:32, Wolfram Schlich [EMAIL PROTECTED] wrote:
So there's no point why the LSB script and the OCF RA could not be one
and the same script, sharing much of the logic.
There are systems that are not using LSB init scripts at all but that
are still used together with software
On 2008-08-20T18:16:41, Robert [EMAIL PROTECTED] wrote:
That's a good idea. Why not put all OCF library stuff into one package
(ocf-base) and add a package for each resource agent (e.g. ocfs-ra-drbd).
Packages are not free. That'd be a mess to maintain. You don't want
that. Trust me. Not
On 2008-08-19T16:40:48, Michael Schwartzkopff [EMAIL PROTECTED] wrote:
Am Dienstag, 19. August 2008 15:55 schrieb Robert:
Hi,
I've been running some tests with quorumd and a three node setup. The
setup is shown in the attached graphics. The cluster uses export
HA_quorum=majority:quorumd.
On 2008-08-20T13:03:50, Adrian Chapela [EMAIL PROTECTED] wrote:
I can't find this bug 1852
(http://developerbugs.linux-foundation.org/show_bug.cgi?id=1852) resolved.
When it could be resolved ?
That was fixed in March and is included, but the commit did not mention
that bugzilla id, because
On 2008-08-19T15:55:13, Robert [EMAIL PROTECTED] wrote:
I played around with hb_setweight - no success (hb_setweight debnode1 200).
The quorum server ignores the weights.
The quorum server is a broken piece of ... code. Avoid it.
Is there a way to force the DC to be in DataCenter1 if
On 2008-08-19T10:47:20, Simon Horman [EMAIL PROTECTED] wrote:
I have uploaded packages for Debian (sid) to debian.org and
they should be available in the debian archive for most architectures
within in the next 24 hours.
Source and i386 binaries are also available at
Hi,
here are my thoughts for 3.0:
- Remove code which depends on pacemaker (mgmt, dopd, CIM/SNMP)
- Remove unmaintained code (telecom/).
- Split up packages so heartbeat's cluster infrastructure layer can be
installed separately from
- supporting libraries, resources, stonith, and lrm code
-
On 2008-08-19T23:13:25, Simon Horman [EMAIL PROTECTED] wrote:
I'm fine with all of this. Though I wonder if it might be best
to get 3.0.0 out the door before worrying about splitting up the packages.
Uhm, but that's the whole point of 3.0.x. ;-)
What release goals for 3.0.0 do you have in
On 2008-08-19T10:47:20, Simon Horman [EMAIL PROTECTED] wrote:
I have uploaded packages for Debian (sid) to debian.org and
they should be available in the debian archive for most architectures
within in the next 24 hours.
Source and i386 binaries are also available at
On 2008-08-18T10:19:56, Andrew Beekhof [EMAIL PROTECTED] wrote:
Fixed in:
http://hg.clusterlabs.org/pacemaker/stable-0.6/rev/2d516888d27c
This is in 2.1.4 as well.
Regards,
Lars
--
Teamlead Kernel, SuSE Labs, Research and Development
SUSE LINUX Products GmbH, GF: Markus Rex, HRB
are available from
http://software.opensuse.org/download/server:/ha-clustering:/lha-2.1
2. The source tar ball available from
http://hg.linux-ha.org/lha-2.1/archive/STABLE-2.1.4.tar.bz2
* Mon Aug 18 2008 Lars Marowsky-Bree [EMAIL PROTECTED] and MANY others (see
doc/AUTHORS)
+ Version
On 2008-08-18T10:30:24, Harakiri [EMAIL PROTECTED] wrote:
Since this is the last 2.1.x Version, i hope that at least for the 3.x
branch, solaris sparc is planned for release.
Sure. We're quite willing to take patches.
3.0.x should be considerably easier to port too, as a lot of code has,
are available from
http://software.opensuse.org/download/server:/ha-clustering:/lha-2.1
2. The source tar ball available from
http://hg.linux-ha.org/lha-2.1/archive/STABLE-2.1.4.tar.bz2
* Mon Aug 18 2008 Lars Marowsky-Bree [EMAIL PROTECTED] and MANY others (see
doc/AUTHORS)
+ Version
On 2008-08-17T22:11:41, Simon Horman [EMAIL PROTECTED] wrote:
I should have made it clear that my target for the work that
I did in debian/ is unstable. If uuid-runtime | uuidgen makes
Etch happy, I'm happy for that change to go in.
It is curious that the packages seem to build for Etch in
On 2008-08-13T10:24:14, Simon Horman [EMAIL PROTECTED] wrote:
Sorry, I've poked around a bit further and I no longer think that
local is POSIX.
People should simply chose whatever language it is they want to write
their scripts in and declare that in the #! line, be it bash, sh, ksh,
csh,
On 2008-08-15T11:55:35, Keisuke MORI [EMAIL PROTECTED] wrote:
I look forward to hearing from Keisuke-san whether this works for them
now!
It does not seem to be fixed right.
It does not cause an assertion failure any more (neither crash ;-),
but an invalid clone resource is appeared.
On 2008-08-15T17:52:42, Keisuke MORI [EMAIL PROTECTED] wrote:
More precisely, we once tried to use clones with 2.1.3 in production
but had to suspend to use it because there were some problems.
Now we want to upgrade it to the coming 2.1.4 with using clones.
_Clones_ by themselves work fine,
,
Fedora. The package version matching this announcement is 2.1.14-17.1.
I will move this to the final location, and update the wiki, and post
a more formal announcement (also cc'ing the announcement list) once the
final version is ready.
* Fri Aug 15 2008 Lars Marowsky-Bree [EMAIL PROTECTED
,
Fedora. The package version matching this announcement is 2.1.14-17.1.
I will move this to the final location, and update the wiki, and post
a more formal announcement (also cc'ing the announcement list) once the
final version is ready.
* Fri Aug 15 2008 Lars Marowsky-Bree [EMAIL PROTECTED
On 2008-08-14T14:24:58, David Avery [EMAIL PROTECTED] wrote:
I originally setup Heartbeat and I setup my haresoruces file and everything
worked fine. After converting my haresources file with the python script and
converting it to my cib.xml file. When I start heartbeat my drbd, mysql, and
On 2008-08-14T10:32:24, Chase Simms [EMAIL PROTECTED] wrote:
I have the following message throughout my logs.
lrmd[3302]: 2008/08/14_08:54:53 info: RA output:
(resource_Pingd:monitor:stderr)
/usr/lib/ocf/resource.d//heartbeat/pingd: line 193: kill: (31237) - No
such process
Please note
On 2008-08-15T13:13:46, Krauth, Alexander [EMAIL PROTECTED] wrote:
Situation:
Running a two node cluster. A master/slave resource with clone_max = 2,
clone_node_max = 1, master_max = 1, master_node_max = 1.
Startup is working, that means 1 clone is running as a master and the
other clone is
On 2008-08-14T16:21:54, Fabricio Vaccari Constanski [EMAIL PROTECTED] wrote:
Hi
I don't speak english veru well, but I will try
I'm using mon and heartbeat.alert script to monitoring apache2 in
debian distro, but not stoping the heartbeat when the service apache2
falls. any suggestions?
On 2008-08-14T23:12:08, Simon Horman [EMAIL PROTECTED] wrote:
Hi Lars, Hi All,
I got to the bottom of why the BasicSanityCheck was haning
on the CRM test. Apparently the CRM test needs access to
the directory in which the log file is stored, and the
new maketempdir code wasn't allowing
On 2008-08-14T16:33:57, Andrew Beekhof [EMAIL PROTECTED] wrote:
But I've got PE crash now when I used with clone resources...
I think the following is the correct fix, but i need to do some more testing
I've pushed that fix for the fatal assert to both the lha-2.1 tree and
the openSUSE build
On 2008-08-13T17:11:54, Keisuke MORI [EMAIL PROTECTED] wrote:
Hi,
I've got an unexpected behavior during our regression test
for the 2.1.4 release.
When the stop of a resource with on_fail=block failed, it looks
like the resource is running on the both nodes according to the
log and
On 2008-08-13T15:01:17, Andrew Beekhof [EMAIL PROTECTED] wrote:
It's fixed in pacemaker (now): 665119a56b2a
Pushed to lha-2.1
___
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page:
On 2008-08-07T09:41:25, David Lee [EMAIL PROTECTED] wrote:
var1=XXX
var2=YYY
my_function() {
_mf_v1=$1
_mf_v2=$2
}
your_function() {
_yf_v1=$1
_yf_v2=$2
}
Pragmatically most sh-programming things are fine with such conventions;
the
On 2008-08-08T17:01:54, Franck Huet [EMAIL PROTECTED] wrote:
heartbeat[7641]: 2008/08/08_16:51:46 WARN: node avil003.sylvea.fr: is dead
heartbeat[10274]: 2008/08/08_16:53:14 WARN: node noeud1: is dead
The two nodes cannot communicate. You need to fix that.
Regards,
Lars
--
Teamlead
On 2008-08-05T18:29:16, Michael Alger [EMAIL PROTECTED] wrote:
It's a bit crude, but it works. The main problem is that if the
monitoring script stops, heartbeat has no idea and therefore can't
factor that into its decision-making process. I haven't determined
whether we can somehow store a
On 2008-07-24T16:29:54, Gerard Petersen [EMAIL PROTECTED] wrote:
The third node is already added to heartbeat config, and in standby mode.
We have contraints in place (full log and config will follow), that work
with the +INF, 'zero' and -INF values, respectively as Master location,
Slave
On 2008-07-31T16:44:31, Angel Rengifo Cancino [EMAIL PROTECTED] wrote:
Yep, it's because I'm first trying to understand very well heartbeat
1.x before learning 2.x style. Using haresources it seems easier for
my simple requirements.
That's not necessarily helpful, as v2 is very different, and
Hi,
sorry, it seems that Fabio hat forgotten to forward this to the
linux-ha-dev list.
Regards,
Lars
--
Teamlead Kernel, SuSE Labs, Research and Development
SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg)
Experience is the name everyone gives to their mistakes. -- Oscar
On 2008-07-22T11:58:31, [EMAIL PROTECTED] wrote:
Hi Mark,
so what you are asking for is a way to externally mark a node down (and
clean/fenced too), because you have additional out-of-band knowledge?
We don't yet have a way of doing this.
A single HBcomm plugin can't deliver, because
On 2008-07-20T17:30:20, Dejan Muhamedagic [EMAIL PROTECTED] wrote:
I'm not sure if we should deal with this at all in heartbeat.
Trying to support all the new setups and distributions, it may
become a nightmare. For the maintenance too: dealing with the
apache configuration is already rather
On 2008-07-20T21:40:12, Florian Knauf [EMAIL PROTECTED] wrote:
From what I can see, the apache RA basically replicates functionality
that's already part of apache2ctl - which is supposed to be used to
start, stop and control apache, anyway. Apache2ctl evaluates the envvars
file if it finds
On 2008-07-21T00:17:04, Choon Kiat [EMAIL PROTECTED] wrote:
Hi,
I running xen virtual machine with heartbeat on. It prompt this
Gmain_timeout_dispatch every 2 or 5 minutes.
The host is not busy as nothing is running currently. Is there any other
ways to optimize heartbeat to prevent this
On 2008-07-18T19:19:39, Dirk H. Schulz [EMAIL PROTECTED] wrote:
What I need is an interface alias with it's own address (like eth0:0)
because of the routing protocol I use at the gateway cluster.
Just out of curiousity, what broken routing protocol requires that?
Oh, it is not broken. It is
On 2008-07-18T16:46:49, Dejan Muhamedagic [EMAIL PROTECTED] wrote:
We are talking about different metrics. All monitor operations
require a fork of a shell.
... which incidentially is also something we eventually need to fix, and
either fork a daemon when the RA starts (which then would talk
On 2008-07-15T14:24:09, Dejan Muhamedagic [EMAIL PROTECTED] wrote:
I'd propose to set the group to the unmanaged mode, do whatever
you have to do to the resource and then put it back to managed:
crm_resource -r grp --meta -p is_managed -v false
stop apache
...
On 2008-07-17T06:49:40, Dirk H. Schulz [EMAIL PROTECTED] wrote:
What I need is an interface alias with it's own address (like eth0:0)
because of the routing protocol I use at the gateway cluster.
Just out of curiousity, what broken routing protocol requires that?
On 2008-07-16T20:42:28, Matthew Soffen [EMAIL PROTECTED] wrote:
The main reason for not wanting bash was for the Non Linux architectures
(i.e. *BSD and Solaris).
Yes, its a readily available but since it isn't standard and sh is, whats
the big deal ?
Desiging/coding for sh (not for bash)
On 2008-07-15T15:18:26, Dejan Muhamedagic [EMAIL PROTECTED] wrote:
There's a tool in Debian which finds bashisms:
checkbashisms(1)
It should find most of bash features. Otherwise, I don't know any
conversion guides. You should be able to find shell references on
Internet. If in doubt,
On 2008-07-16T13:56:27, HIDEO YAMAUCHI [EMAIL PROTECTED] wrote:
Hi All,
The resource of Xen uses 'xm destroy -w .'.
But, this command does not have the w option.
I think that it is a typo.
Which Xen version are you running on?
And yeah, more recent versions seem to have cut that
On 2008-07-16T11:19:53, Schmidt, Florian [EMAIL PROTECTED] wrote:
What can I do?
Provide more information in the form of your CIB and the actual version
of our heartbeat and drbd you're running, for a start ;-)
Regards,
Lars
--
Teamlead Kernel, SuSE Labs, Research and Development
SUSE
On 2008-07-14T21:20:10, HIDEO YAMAUCHI [EMAIL PROTECTED] wrote:
Hi Lars,
I'm sorry.
I am not good at English. Regards,
But yes, it appears that the watchdog timeout should be less than
deadtime.
However, the idea and I of you are the same.
Can you reflect this demand?
Or, should I
On 2008-07-14T10:13:21, Nikhil Kulkarni [EMAIL PROTECTED] wrote:
This is my haresources file:
watchdog-client1 IPaddr::10.0.38.71/24/eth0 drbddisk::r0 Delay::3::0
Filesystem::/dev/drbd1:://mnt/data::ext3 kill nfs Delay::3::0 nfs
nfslock
Why all the Delay resources? They don't help
On 2008-07-14T19:25:30, Bala [EMAIL PROTECTED] wrote:
Jul 12 01:27:43 w2k8-src heartbeat: [5424]: ERROR: MSG[3] : [protocol=1]
My /etc/ha.d/ha.cnf has the following two nodes defined in the HA
configuration which match with the uname -n output as recommended:
On 2008-07-15T08:13:25, Paul Walsh [EMAIL PROTECTED] wrote:
I have the following resource group defined:
Resource Group: Moodle
web_dev (heartbeat:drbddisk): Started mercury
mysql_dev (heartbeat:drbddisk): Started mercury
weblog_dev(heartbeat:drbddisk): Started
On 2008-07-15T10:56:44, Paul Walsh [EMAIL PROTECTED] wrote:
/usr/lib/ocf/resource.d/BCU/apache2 stop
/usr/lib/ocf/resource.d/BCU/apache2 start
No, not if you're doing monitoring; the cluster will find out and
restart the group.
The group, or just the resource? In theory, the script should
On 2008-07-15T15:54:10, Bala [EMAIL PROTECTED] wrote:
Do you have several clusters on the same network segment? If so, you
should put them on different port numbers (udpport) or multicast
addresses.
Actually, yes. I do have a different cluster on the same network.
Thanks for the pointer. I
On 2008-07-15T16:19:31, HIDEO YAMAUCHI [EMAIL PROTECTED] wrote:
Hi Lars,
I forgot to say a very important thing.
The system which I intend for does not use STONITH.
If STONITH enters, I do not become such a problem.
Systems without STONITH are not supportable anway.
Regards,
Lars
On 2008-07-14T08:05:52, Ciro Iriarte [EMAIL PROTECTED] wrote:
Checking the Xen agent I see that the configuration file is required.
Can't xen sync the configuration by itself at migration time
(xend-relocation-server)?, I case this is false, I still need to
export and sync the
On 2008-07-14T12:27:58, HIDEO YAMAUCHI [EMAIL PROTECTED] wrote:
Hi All,
I confirmed it about watchdog of Heartbeat.
The environment that I confirmed has two resources in two
nodes(active/standby).
The resource of one eyes checks starting VIP.
The resource of two eyes is VIP(IPaddr).
On 2008-07-12T00:06:10, Ciro Iriarte [EMAIL PROTECTED] wrote:
Hi, currently Xen 3.2 on SLES10 managed domains ('xm new' instead of
'xm create') only uses configuration files the first time you add
them, later, any 'xm block-attach' is stored in its own database (not
sure about the
601 - 700 of 1358 matches
Mail list logo
