[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-ha-dev
Subject:    [Linux-ha-dev] Re: SUSE 9.2 64 Bit and Heartbeat 1.99-4
From:       Andrew Beekhof (GMail) <beekhof () gmail ! com>
Date:       2005-04-15 8:08:33
Message-ID: 60476ef1976f2de3f3253c8f47c39e7d () gmail ! com
[Download RAW message or body]


On Apr 14, 2005, at 8:25 PM, Alexander Schanz wrote:

> Andrew Beekhof (GMail) wrote:
>
>>
>> On Apr 14, 2005, at 4:22 PM, Alan Robertson wrote:
>>
>>> Alexander Schanz wrote:
>>>
>>>> Hello Alan,
>>>> I have written a rpm.spec file to get heartbeat running on Suse 9.2 
>>>> 64Bit (AMD) and 32 Bit. I have used a BUILDROOT to make the whole 
>>>> process better controlable. If you are interested I can send you 
>>>> the *.spec file for inclusion in the cvs.
>>>
>>>
>>> That's what I build with all the time. My desktop machine is an 
>>> AMD64 running SUSE 9.2. The spec files we generate from CVS build 
>>> for all those platforms. I just don't usually bother putting up the 
>>> AMD64 RPMs for non-production releases.
>>
> I have seen your *.spec file but, I have added some extras like 
> variables for the different options ( building the snmp-agent for 
> example). I addition to this we usually use variables like 'libdir' to 
> make the whole *.spec file more generic. The build dependencies have 
> also been added to match SUSE 9.2 and we will create an other *.spec 
> for 9.3 tomorrow.
> Oh and before I forget I got hbagent running in the end – maybe 
> someone should mention that a 'export MIBS = + LINUX-HA-MIB' before 
> starting the snmpd is a good idea
>
>>>
>>>> The problem I have now is that I would like to have a nice example 
>>>> on how to add resources and resourcegroups to the new cib. We are 
>>>> testing the new version since we will need the ability to transfer 
>>>> resources between cluster members. the provides test-sripts are not 
>>>> very helpful.
>>>
>>>
>>> We're concentrating our energies on testing the code and getting it 
>>> right at this point.
>>>
>>> But, I'm sure Andrew would be happy to answer questions - especially 
>>> if you'd document the results of your discussions ;-).
>>
>>
>> You bet. What have you read so far and what problems have you been 
>> experiencing?
>>
> Let me explain what an why we want to test 1.99-4 in the first place.
>
> I'm the Head of the Linux Competence and Service Centre of the German 
> Air Navigation Services (www.dfs.de).
>
> (As the manager I have the luxury to decide which projects I want to 
> be involved with ... therefore I decided that clustering is 
> interesting.)
>
> We are in the process of converting many operational Systems to Linux. 
> One of the systems is currently using HP-UX and HP's ClusterManger. 
> One of the requirements is to be able to shift groups of resource from 
> one cluster member to an other. Since this is not possible with 1.2.3 
> we are currently testing 1.99-4. The port of the system has to be 
> finished by July this year an we have to decide within the next weeks 
> whether to use 1.99-4 , fall back to 1.2.3 (and live with the 
> unsatisfied requirement until 2.0 is out and stable) or use a 
> commercial product.

Without wanting to deter you in anyway from putting 1.99.4 through its 
paces (and I'd love to hear feedback), think of the 1.99.x series as an 
evaluation series for 2.0.  1.99.x should not be used in production.  
Ever.

2.0.0, when it comes out, will have passed much higher quality 
requirements and essentially be the bug-fixed culmination of 1.99.x

In terms of when that will be, that is Alan's call, but I can say that 
we are now feature complete and are working through the exhaustive 
process of testing the hell out of it.  At the moment the CRM has 5 
blocker-bugs, of which one will take about 5 minutes, 1 is probably 
syslog acting up, and 2 are more about working optimally rather than 
doing the wrong thing... so we're closer than ever :)

>
> We have tested 1.2.3 so far in combination with DRBD and MONIT.
>
> 1.99-4 with hbagent,crm,cib is running – without resources...
>
> For 1.99-4 I would like to know how define the resources and group 
> them.
> Is there an add-resource script or do I really have to edit the XLM 
> files for this?

A GUI is planned (I believe the person is nearly done with a preview), 
but for now one must either edit the XML file, or create XML fragments.

Unfortunately this is a short-term necessary evil... the flexibility 
that 2.x will provides makes writing admin tools exceedingly 
non-trivial - there are just so many combinations and possibilities.

What we can do is attempt to make that as painless as possible.

> Any examples to follow ?
>

Like everything else at the moment, examples are "on the list" and are 
mostly scattered through the various testing pieces we have.

I might try and pull a few things together today... I need a break from 
coding and the sun is out so its a good day to work outside :)

>
>
>>>
>>> Release 1 can transfer resources between cluster members - but it 
>>> cannot move individual resources or resource groups around. In 
>>> haresources whatever you put as belonging to one machine can be 
>>> moved around as a whole.
>>>
>>> But with release 2, you can constrain where things run, and you can 
>>> change the constraints on any individual resource at run time. When 
>>> you change the constraints, it will then act on them - moving that 
>>> resource (and those that depend on it) to whereever the constraints 
>>> dictate that it should go.
>>>
>>>
>>> -- 
>>> Alan Robertson <alanr@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
>>
>> "No means no, and no means yes, and everything in between and all the 
>> rest" - TISM
>>
>
> Did I mention that we are currently looking for Linux Kernel 
> developers and other Linux and C specialists?
>
> --
> Alexander Schanz
>
>
--
Andrew Beekhof

"If it weren't for my horse, I wouldn't have spent that year in 
college" - Unknown, courtesy of Lewis Black

_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/

[prev in list] [next in list] [prev in thread] [next in thread]