Mailing List Archive: Linux-HA: Users

Announcing! Release 2.0.0 of Linux-HA is now available!

Index | Next | Previous | View Threaded

alanr at unix

Jul 29, 2005, 9:19 PM

Post #1 of 14 (5845 views)
Permalink

Announcing! Release 2.0.0 of Linux-HA is now available!

The Linux-HA team proudly announces the long-anticipated release 2.0.0
of the Linux-HA (aka "heartbeat") software.

You can find it here:
http://linux-ha.org/download/index.html#2.0.0

This release extends the capabilities of Linux-HA far beyond anything
available in the past, and provides basic capabilities comparable to
any commercial HA package.

This release provides support for monitoring of resources (services) and
support for larger clusters - we have tested up to 16-node clusters.
In Release 2, simple clusters are simple to create, and more complex
clusters can take advantage of our rule-based resource placement methods
to ensure that the cluster does exactly what is desired when failures occur.

In addition, it supports the OCF standard resource model, the SAF
membership API, and provides command line, web-based and SNMP-based
cluster monitoring tools.

It extends our tradition of high quality through excellent code,
exhaustive automated testing, zero warnings in source, and for the first
time, all project code passes a static analyzer with no complaints.

Many thanks to the whole crew of people who worked so hard to bring you
this release! (i.e, they wrote most of the new code).

Change Log:

* Fri 29 July 2005 Alan Robertson <alanr[at]unix.sh> (see doc/AUTHORS file)
+ Version 2.0.0 - First stable release of the next generation of the
Linux-HA project
+ Basic Characteristics described here:
http://linux-ha.org/FactSheetv2
+ Core infrastructure improvements:
+ Messaging (message acknowledging and flow control)
+ Logging (logging daemon)
+ Release 1.x style (2-node) clusters fully supported
+ Multi-node support (so far up to 16-node clusters tested)
See http://linux-ha.org/GettingStartedV2 for more information
+ New components:
+ Cluster Information Base (replicated resource configuration)
+ Cluster Resource Manager (supporting 1->N nodes)
+ Modular Policy Engine (controlling resource placement)
+ Local Resource Manager (policy free, not cluster aware)
+ Stonith Daemon (stand-alone fencing subsystem)
+ Support for OCF and LSB resource agents
+ Support for composite resource types (groups, clones)
+ Support for a rich set of resource location and ordering constraints
+ Conversion tool for existing haresources
+ Resources monitored by request
+ Resource "maintenance" mode
+ Several failback, failure and "No Quorum" behaviors to choose from
(global defaults and per action or resource)
+ Sample cluster state and configuration monitoring tools

Known issues in 2.0.0:
- Under some rare circumstances the cluster manager will time out
while stabilizing a new cluster state. This appears to be
otherwise harmless - the cluster is actually fine.
http://www.osdl.org/developer_bugzilla/show_bug.cgi?id=770
- Under some rare circumstances, a dev assert will be triggered
in unpack.c. This results in various things getting restarted.
This is annoying, but is recovered from correctly.
http://www.osdl.org/developer_bugzilla/show_bug.cgi?id=797

--
Alan Robertson <alanr[at]unix.sh>

"Openness is the foundation and preservative of friendship... Let me
claim from you at all times your undisguised opinions." - William
Wilberforce
_______________________________________________
Linux-HA mailing list
Linux-HA[at]lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha

horms at verge

Jul 30, 2005, 4:01 AM

Post #2 of 14 (5694 views)
Permalink

Re: [Linux-ha-dev] Announcing! Release 2.0.0 of Linux-HA is now available! [In reply to]

On Fri, Jul 29, 2005 at 10:19:27PM -0600, Alan Robertson wrote:
> The Linux-HA team proudly announces the long-anticipated release 2.0.0
> of the Linux-HA (aka "heartbeat") software.
>
> You can find it here:
> http://linux-ha.org/download/index.html#2.0.0

I have added debian packages for 2.0.0 to
http://www.ultramonkey.org/download/heartbeat/2.0.0/

I have also uploaded these to Debian, so hopefully
they will appear there soon. I have made these
packages hearteat-2, so the old 1.2.3 can remain
in Debian as heartbeat for now. This should allow
people more flexibility to choose which version
of heartbeat they want to run, though envisage
only haveing a 2.0.x package in Debian in the
not to distant future.

--
Horms
_______________________________________________
Linux-HA mailing list
Linux-HA[at]lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha

lmb at suse

Jul 30, 2005, 8:27 AM

Post #3 of 14 (5689 views)
Permalink

Re: [Linux-ha-dev] Announcing! Release 2.0.0 of Linux-HA is now available! [In reply to]

On 2005-07-29T22:19:27, Alan Robertson <alanr[at]unix.sh> wrote:

Three cheers!

In particular to Alan, Andrew and IBM Bejing who carried the bulk of the
load this time.

SLES9 packages will be made available on Monday.

Sincerely,
Lars Marowsky-Brée <lmb[at]suse.de>

--
High Availability & Clustering
SUSE Labs, Research and Development
SUSE LINUX Products GmbH - A Novell Business -- Charles Darwin
"Ignorance more frequently begets confidence than does knowledge"

_______________________________________________
Linux-HA mailing list
Linux-HA[at]lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha

alanr at unix

Jul 30, 2005, 9:47 AM

Post #4 of 14 (5702 views)
Permalink

Re: [Linux-ha-dev] Announcing! Release 2.0.0 of Linux-HA is now available! [In reply to]

Lars Marowsky-Bree wrote:
> On 2005-07-29T22:19:27, Alan Robertson <alanr[at]unix.sh> wrote:
>
> Three cheers!
>
> In particular to Alan, Andrew and IBM Bejing who carried the bulk of the
> load this time.

Andrew, Sun Jiang Dong, Huang Zhen and Pan Jia Ming did the heavy
lifting here - not me. I mainly stood around and criticized ;-).

> SLES9 packages will be made available on Monday.

Of course, if you don't want to wait 'til Monday, you should feel free
to use the SUSE 9.2/9.3 professional packages already on the web site.
They'll likely work quite nicely for you.

--
Alan Robertson <alanr[at]unix.sh>

"Openness is the foundation and preservative of friendship... Let me
claim from you at all times your undisguised opinions." - William
Wilberforce
_______________________________________________
Linux-HA mailing list
Linux-HA[at]lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha

lmb at suse

Aug 1, 2005, 6:32 AM

Post #5 of 14 (5711 views)
Permalink

Re: Announcing! Release 2.0.0 of Linux-HA is now available! [In reply to]

On 2005-07-29T22:19:27, Alan Robertson <alanr[at]unix.sh> wrote:

> The Linux-HA team proudly announces the long-anticipated release 2.0.0
> of the Linux-HA (aka "heartbeat") software.
>
> You can find it here:
> http://linux-ha.org/download/index.html#2.0.0

Packages for SLES9 testing have been placed at
ftp://ftp.suse.com/pub/people/lmb/heartbeat/.

These will track CVS over time now. Please see the README there too ;-)

Sincerely,
Lars Marowsky-Brée <lmb[at]suse.de>

--
High Availability & Clustering
SUSE Labs, Research and Development
SUSE LINUX Products GmbH - A Novell Business -- Charles Darwin
"Ignorance more frequently begets confidence than does knowledge"

_______________________________________________
Linux-HA mailing list
Linux-HA[at]lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha

pmueller at sidestep

Aug 1, 2005, 10:54 AM

Post #6 of 14 (5697 views)
Permalink

RE: Announcing! Release 2.0.0 of Linux-HA is now available! [In reply to]

> The Linux-HA team proudly announces the long-anticipated
> release 2.0.0 of the Linux-HA (aka "heartbeat") software.
>
> You can find it here:
> http://linux-ha.org/download/index.html#2.0.0
>
> This release extends the capabilities of Linux-HA far beyond
> anything available in the past, and provides basic
> capabilities comparable to any commercial HA package.

Thanks a lot HA-team, big changes in 2.0! Now to grab my copy for testing...

One question going in : are there legacy differences from 1.x -> 2.0 that
would cause an upgrade to break? These are probably in README/INSTALL
somewhere, but if you use an RPM it is sometimes difficult to get this
information.

Regards,

P
_______________________________________________
Linux-HA mailing list
Linux-HA[at]lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha

alanr at unix

Aug 2, 2005, 3:20 AM

Post #7 of 14 (5688 views)
Permalink

Re: Announcing! Release 2.0.0 of Linux-HA is now available! [In reply to]

Peter Mueller wrote:
>>The Linux-HA team proudly announces the long-anticipated
>>release 2.0.0 of the Linux-HA (aka "heartbeat") software.
>>
>>You can find it here:
>> http://linux-ha.org/download/index.html#2.0.0
>>
>>This release extends the capabilities of Linux-HA far beyond
>>anything available in the past, and provides basic
>>capabilities comparable to any commercial HA package.
>
> Thanks a lot HA-team, big changes in 2.0! Now to grab my copy for testing...
>
> One question going in : are there legacy differences from 1.x -> 2.0 that
> would cause an upgrade to break? These are probably in README/INSTALL
> somewhere, but if you use an RPM it is sometimes difficult to get this
> information.

If you install a legacy (1.x) style cluster, there shouldn't be anything
like this.

If you upgrade to a CRM-managed cluster, there are lots of things which
need to be taken care of. Andrew has documented how to convert your
haresources file over to R2, and we even provided a tool (probably
undocumented).

I'm not sure if we've documented a live cutover procedure for this
conversion to a CRM-managed (r2-style) cluster yet. Andrew?

--
Alan Robertson <alanr[at]unix.sh>

"Openness is the foundation and preservative of friendship... Let me
claim from you at all times your undisguised opinions." - William
Wilberforce
_______________________________________________
Linux-HA mailing list
Linux-HA[at]lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha

beekhof at gmail

Aug 2, 2005, 4:09 AM

Post #8 of 14 (5705 views)
Permalink

Re: Announcing! Release 2.0.0 of Linux-HA is now available! [In reply to]

On Aug 2, 2005, at 12:20 PM, Alan Robertson wrote:

> Peter Mueller wrote:
>
>>> The Linux-HA team proudly announces the long-anticipated release
>>> 2.0.0 of the Linux-HA (aka "heartbeat") software.
>>>
>>> You can find it here:
>>> http://linux-ha.org/download/index.html#2.0.0
>>>
>>> This release extends the capabilities of Linux-HA far beyond
>>> anything available in the past, and provides basic capabilities
>>> comparable to any commercial HA package.
>>>
>> Thanks a lot HA-team, big changes in 2.0! Now to grab my copy for
>> testing...
>> One question going in : are there legacy differences from 1.x ->
>> 2.0 that
>> would cause an upgrade to break? These are probably in README/
>> INSTALL
>> somewhere, but if you use an RPM it is sometimes difficult to get
>> this
>> information.
>>
>
> If you install a legacy (1.x) style cluster, there shouldn't be
> anything like this.
>
> If you upgrade to a CRM-managed cluster, there are lots of things
> which need to be taken care of. Andrew has documented how to
> convert your haresources file over to R2, and we even provided a
> tool (probably undocumented).
>
> I'm not sure if we've documented a live cutover procedure for this
> conversion to a CRM-managed (r2-style) cluster yet. Andrew?
>

not yet - but the short, unapproved, off the top of my head version
would be...

1) shutdown heartbeat on node B (so all the resources move to A)
2) kill -9 heartbeat on node A (so the resources are still running)
3) upgrade heartbeat on both nodes
4) reconfigure with v2
5) make sure your resource scripts treat starting a started resource
as a success case
6) bring heartbeat on node A up
7) wait a bit
8) check the node is stable ("crmadmin -s nodeA" should return a
string with S_IDLE in it)
9) start heartbeat on node B

not pretty...

>
> --
> Alan Robertson <alanr[at]unix.sh>
>
> "Openness is the foundation and preservative of friendship... Let
> me claim from you at all times your undisguised opinions." -
> William Wilberforce
>

--
Andrew Beekhof

"Too much knowledge leads to confusion; Too many guitar lessons lead
to jazz-fusion!" - TISM

_______________________________________________
Linux-HA mailing list
Linux-HA[at]lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha

lmb at suse

Aug 2, 2005, 4:23 AM

Post #9 of 14 (5708 views)
Permalink

Re: Announcing! Release 2.0.0 of Linux-HA is now available! [In reply to]

On 2005-08-02T13:09:12, "Andrew Beekhof (GMail)" <beekhof[at]gmail.com> wrote:

> not yet - but the short, unapproved, off the top of my head version
> would be...
>
> 1) shutdown heartbeat on node B (so all the resources move to A)
> 2) kill -9 heartbeat on node A (so the resources are still running)
> 3) upgrade heartbeat on both nodes
> 4) reconfigure with v2
> 5) make sure your resource scripts treat starting a started resource
> as a success case
> 6) bring heartbeat on node A up

Just out of curiousity, don't we check whether the resources are running
when we start, and not start resources which are already running?

Certainly you're right that start must be idempotent, but I was just
wondering whether we actually do rely on that in this specific scenario.

> not pretty...

Indeed it isn't. I'd much rather setup a new cluster in the lab and flip
the switch on the whole cluster after testing.

(If one is short on machines, I'd split one node off the existing
cluster, migrate and test that, and then bring the legacy one down and
add it to that new cluster.)

Sincerely,
Lars Marowsky-Brée <lmb[at]suse.de>

--
High Availability & Clustering
SUSE Labs, Research and Development
SUSE LINUX Products GmbH - A Novell Business -- Charles Darwin
"Ignorance more frequently begets confidence than does knowledge"

_______________________________________________
Linux-HA mailing list
Linux-HA[at]lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha

beekhof at gmail

Aug 2, 2005, 4:37 AM

Post #10 of 14 (5701 views)
Permalink

Re: Announcing! Release 2.0.0 of Linux-HA is now available! [In reply to]

On Aug 2, 2005, at 1:23 PM, Lars Marowsky-Bree wrote:

> On 2005-08-02T13:09:12, "Andrew Beekhof (GMail)"
> <beekhof[at]gmail.com> wrote:
>
>
>> not yet - but the short, unapproved, off the top of my head version
>> would be...
>>
>> 1) shutdown heartbeat on node B (so all the resources move to A)
>> 2) kill -9 heartbeat on node A (so the resources are still running)
>> 3) upgrade heartbeat on both nodes
>> 4) reconfigure with v2
>> 5) make sure your resource scripts treat starting a started resource
>> as a success case
>> 6) bring heartbeat on node A up
>>
>
> Just out of curiousity, don't we check whether the resources are
> running
> when we start, and not start resources which are already running?
>
> Certainly you're right that start must be idempotent, but I was just
> wondering whether we actually do rely on that in this specific
> scenario.

we discussed starting monitors for all resources on all machines but
i thought we decided against it (at least for now)

>
>
>> not pretty...
>>
>
> Indeed it isn't. I'd much rather setup a new cluster in the lab and
> flip
> the switch on the whole cluster after testing.
>
> (If one is short on machines, I'd split one node off the existing
> cluster, migrate and test that, and then bring the legacy one down and
> add it to that new cluster.)
>
>
> Sincerely,
> Lars Marowsky-Brée <lmb[at]suse.de>
>
> --
> High Availability & Clustering
> SUSE Labs, Research and Development
> SUSE LINUX Products GmbH - A Novell Business -- Charles Darwin
> "Ignorance more frequently begets confidence than does knowledge"
>
> _______________________________________________
> Linux-HA mailing list
> Linux-HA[at]lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
>

--
Andrew Beekhof

"Eating fruit is mean and vicious... keep your hands off Golden
Delicious" - TISM

_______________________________________________
Linux-HA mailing list
Linux-HA[at]lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha

lmb at suse

Aug 2, 2005, 4:39 AM

Post #11 of 14 (5708 views)
Permalink

Re: Announcing! Release 2.0.0 of Linux-HA is now available! [In reply to]

On 2005-08-02T13:37:34, "Andrew Beekhof (GMail)" <beekhof[at]gmail.com> wrote:

> >Certainly you're right that start must be idempotent, but I was just
> >wondering whether we actually do rely on that in this specific
> >scenario.
> we discussed starting monitors for all resources on all machines but
> i thought we decided against it (at least for now)

Uhm. We must do that to detect whether a resource is running somewhere
where we don't think it should be running! We _MUST_! That is like one
of the top three error admins make - starting the resources somewhere on
their own.

And if we don't detect this, we might be responsible for starting it
somewhere else. _Twice_. Which we MUST NOT do.

Sincerely,
Lars Marowsky-Brée <lmb[at]suse.de>

--
High Availability & Clustering
SUSE Labs, Research and Development
SUSE LINUX Products GmbH - A Novell Business -- Charles Darwin
"Ignorance more frequently begets confidence than does knowledge"

_______________________________________________
Linux-HA mailing list
Linux-HA[at]lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha

bernhard.limbach at comsoft

Aug 2, 2005, 5:52 AM

Post #12 of 14 (5702 views)
Permalink

Re: Announcing! Release 2.0.0 of Linux-HA is now available! [In reply to]

Hello,

just two little questions when trying to get started with 2.0.0:

1. What is the official way to set up an initial CIB ?

From the documentation and some comments in the mailing lists I
deduced that I have to use cibadmin (exclusivly?) to modify the CIB.
As far as I understand, this only works when heartbeat is running
already.
Thus I started heartbeat on all nodes with "crm yes" in ha.cf and it
started without resources and assumedly some default settings.
Then I added a resource and the resource was started immediately on
one of the nodes.

Well, I could go on like this, but what I am missing is some way to
prepare the system in advance without having heartbeat running. (The
old way with 1.2.3 was to edit ha.cf and haresources and then
starting heartbeat, wasn't it ?)

2. What is the implementation status of multi state resources (master/
slave) (especially drbd) ?

I found some drafts in the documentation and the OCF drbd resource
agent shipped with 2.0.0 contains the promote/demote commands but I
could not find out whether it is recommended to use this or the old
way (heartbeat-type drbddisk resource agent).

Just a little confused about some basic ideas...

Sincerly,
Bernhard Limbach

--
Dr. Bernhard LIMBACH Phone: +49 721 9497 - 241
COMSOFT GmbH Fax: +49 721 9497 - 129
Wachhausstrasse 5a http://www.comsoft.de
76227 Karlsruhe, Germany mailto: bernhard.limbach[at]comsoft.de

_______________________________________________
Linux-HA mailing list
Linux-HA[at]lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha

beekhof at gmail

Aug 2, 2005, 6:21 AM

Post #13 of 14 (5697 views)
Permalink

Re: Announcing! Release 2.0.0 of Linux-HA is now available! [In reply to]

On Aug 2, 2005, at 2:52 PM, Bernhard Limbach wrote:

> Hello,
>
> just two little questions when trying to get started with 2.0.0:
>
> 1. What is the official way to set up an initial CIB ?
>
> From the documentation and some comments in the mailing lists I
> deduced that I have to use cibadmin (exclusivly?) to modify the CIB.
> As far as I understand, this only works when heartbeat is running
> already.

correct

> Thus I started heartbeat on all nodes with "crm yes" in ha.cf and it
> started without resources and assumedly some default settings.
> Then I added a resource and the resource was started immediately on
> one of the nodes.
>
> Well, I could go on like this, but what I am missing is some way to
> prepare the system in advance without having heartbeat running. (The
> old way with 1.2.3 was to edit ha.cf and haresources and then
> starting heartbeat, wasn't it ?)

If every node in the cluster is stopped, only then can you edit the
CIB by hand.
The node on which you made the changes should then be started first.
Is is also a good idea to increment admin_epoch when doing this.

alternatively, you can add the resources as being initially unmanaged
(http://www.linux-ha.org/ClusterResourceManager/DTD1.0/
Annotated#is_managed)

and only set them to be managed once you have put in the rest of the
configuration.

>
> 2. What is the implementation status of multi state resources (master/
> slave) (especially drbd) ?
>

one of the two highest priority features that will be in the upcoming
point releases (2.0.x).

> I found some drafts in the documentation and the OCF drbd resource
> agent shipped with 2.0.0 contains the promote/demote commands but I
> could not find out whether it is recommended to use this or the old
> way (heartbeat-type drbddisk resource agent).
>
>
> Just a little confused about some basic ideas...
>
> Sincerly,
> Bernhard Limbach
>
> --
> Dr. Bernhard LIMBACH Phone: +49 721 9497 - 241
> COMSOFT GmbH Fax: +49 721 9497 - 129
> Wachhausstrasse 5a http://www.comsoft.de
> 76227 Karlsruhe, Germany mailto: bernhard.limbach[at]comsoft.de
>
>
> _______________________________________________
> Linux-HA mailing list
> Linux-HA[at]lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
>

--
Andrew Beekhof

"Would the last person to leave please turn out the enlightenment?" -
TISM

_______________________________________________
Linux-HA mailing list
Linux-HA[at]lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha

lmb at suse

Aug 2, 2005, 6:22 AM

Post #14 of 14 (5700 views)
Permalink

Re: Announcing! Release 2.0.0 of Linux-HA is now available! [In reply to]

On 2005-08-02T14:52:58, Bernhard Limbach <bernhard.limbach[at]comsoft.de> wrote:

> Well, I could go on like this, but what I am missing is some way to
> prepare the system in advance without having heartbeat running. (The
> old way with 1.2.3 was to edit ha.cf and haresources and then
> starting heartbeat, wasn't it ?)

You can put a whole CIB in place via crmadmin in one go, instead of
adding resources one per one. But yes, this will only work with
heartbeat/CRM running.

You can put a CIB into /var/lib/heartbeat/crm/cib.xml prior to startup,
but after startup, you may only use the tool interfaces and not modify
the file directly.

> 2. What is the implementation status of multi state resources (master/
> slave) (especially drbd) ?

Not started yet.

Sincerely,
Lars Marowsky-Brée <lmb[at]suse.de>

--
High Availability & Clustering
SUSE Labs, Research and Development
SUSE LINUX Products GmbH - A Novell Business -- Charles Darwin
"Ignorance more frequently begets confidence than does knowledge"

_______________________________________________
Linux-HA mailing list
Linux-HA[at]lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha

Index | Next | Previous | View Threaded

Interested in having your list archived? Contact lists@gossamer-threads.com