Login | Register For Free | Help
Search for: (Advanced)

Mailing List Archive: Linux-HA: Users

active/active cluster

 

 

Linux-HA users RSS feed   Index | Next | Previous | View Threaded


sylv1clerc at gmail

May 3, 2005, 12:39 AM

Post #1 of 11 (1321 views)
Permalink
active/active cluster

Hello,

I'm sorry I think the question has already been asked but I don't
understand very well how I can do.

I have a freeradius server (so I needn't a shared storage solution)
but I work in a big enterprise and I must do an active/active server
to decrease the charge of the processor.

For my test, I do an active/passive server which work fine but I don't
understand how do an active/active server with hearbeat only for 2
daemon (freeradius & samba). I only need that the requests which
arrive to the server are shared between the 2 servers.

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


kenneth at geisshirt

May 3, 2005, 4:31 AM

Post #2 of 11 (1309 views)
Permalink
Re: active/active cluster [In reply to]

Sylvain Clerc wrote:

> For my test, I do an active/passive server which work fine but I don't
> understand how do an active/active server with hearbeat only for 2
> daemon (freeradius & samba). I only need that the requests which
> arrive to the server are shared between the 2 servers.

I'm not sure that I follow you but I give it a shot. In an active/active
configuration you have both nodes running some of the services/daemons.
Which node is the prefered one is specified in the file
/etc/ha.d/haresources by having the init script on seperate line as in
this example (not including shared storage):

node_A 172.16.0.11 freeradius
node_B 172.16.0.12 samba

node_A will have IP address 172.16.0.11 and run freeradius while node_B
will have IP address 172.16.0.12 and run Samba. Your Windows clients
will connect to address 172.16.0.12, and if node_B fails, node_A will
take over address 172.16.0.12 and start Samba.

In real life you'll need shared storage for Samba (DRDB might be
useful). In this case in particular you MUST install and configure some
king of STONITH (remote power cycle) device.

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


lmb at suse

May 3, 2005, 4:53 AM

Post #3 of 11 (1318 views)
Permalink
Re: active/active cluster [In reply to]

On 2005-05-03T13:31:12, Kenneth Geisshirt <kenneth [at] geisshirt> wrote:

> node_A 172.16.0.11 freeradius
> node_B 172.16.0.12 samba

... except that an underscore is not valid in a nodename. Just thought
I'd point that out ;-)


--
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
http://lists.linux-ha.org/mailman/listinfo/linux-ha


sylv1clerc at gmail

May 3, 2005, 5:30 AM

Post #4 of 11 (1294 views)
Permalink
Re: active/active cluster [In reply to]

> I'm not sure that I follow you but I give it a shot. In an active/active
> configuration you have both nodes running some of the services/daemons.
> Which node is the prefered one is specified in the file
> /etc/ha.d/haresources by having the init script on seperate line as in
> this example (not including shared storage):

In fact, I would the 2servers run the 2 services and share the
resources but I read on the web that it's not possible only with
hearbeat and it needs another computer to do that. As I want to use
only my 2servers, I give up that. (Tell me if I'm wrong)

> In real life you'll need shared storage for Samba (DRDB might be
> useful). In this case in particular you MUST install and configure some
> king of STONITH (remote power cycle) device.

I use Samba only to join an Active Directory with freeradius, I
needn't to configure it (exept precise the domain name), the service
just have to run.

Thank you for your response,
Sylvain Clerc.
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha


kenneth at geisshirt

May 3, 2005, 5:39 AM

Post #5 of 11 (1283 views)
Permalink
Re: active/active cluster [In reply to]

Sylvain Clerc wrote:
>>I'm not sure that I follow you but I give it a shot. In an active/active
>>configuration you have both nodes running some of the services/daemons.
>>Which node is the prefered one is specified in the file
>>/etc/ha.d/haresources by having the init script on seperate line as in
>>this example (not including shared storage):
>
>
> In fact, I would the 2servers run the 2 services and share the
> resources but I read on the web that it's not possible only with
> hearbeat and it needs another computer to do that. As I want to use
> only my 2servers, I give up that. (Tell me if I'm wrong)

You use Heartbeat to control which of the two server that runs the
services. Heartbeat 1.2.x only supports 2-server solutions.

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


lmb at suse

May 3, 2005, 5:44 AM

Post #6 of 11 (1309 views)
Permalink
Re: active/active cluster [In reply to]

On 2005-05-03T14:30:45, Sylvain Clerc <sylv1clerc [at] gmail> wrote:

> In fact, I would the 2servers run the 2 services and share the
> resources but I read on the web that it's not possible only with
> hearbeat and it needs another computer to do that. As I want to use
> only my 2servers, I give up that. (Tell me if I'm wrong)

That question doesn't make sense. What do you want to do?


Mit freundlichen Grüßen,
Lars Marowsky-Brée <lmb [at] suse>

--
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
http://lists.linux-ha.org/mailman/listinfo/linux-ha


alanr at unix

May 3, 2005, 6:20 AM

Post #7 of 11 (1307 views)
Permalink
Re: active/active cluster [In reply to]

Sylvain Clerc wrote:
> Hello,
>
> I'm sorry I think the question has already been asked but I don't
> understand very well how I can do.
>
> I have a freeradius server (so I needn't a shared storage solution)
> but I work in a big enterprise and I must do an active/active server
> to decrease the charge of the processor.
>
> For my test, I do an active/passive server which work fine but I don't
> understand how do an active/active server with hearbeat only for 2
> daemon (freeradius & samba). I only need that the requests which
> arrive to the server are shared between the 2 servers.


This page has two active/active configurations illustrated:
http://wwnew.linux-ha.org/GettingStartedWithHeartbeat

See Also:
http://wwnew.linux-ha.org/ActiveActive

--
Alan Robertson <alanr [at] unix>

"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
http://lists.linux-ha.org/mailman/listinfo/linux-ha


jdm at operamail

May 4, 2005, 2:34 AM

Post #8 of 11 (1318 views)
Permalink
Re: active/active cluster [In reply to]

Sylvain Clerc wrote:

>I have a freeradius server (so I needn't a shared storage solution)
>but I work in a big enterprise and I must do an active/active server
>to decrease the charge of the processor.
>
>For my test, I do an active/passive server which work fine but I don't
>understand how do an active/active server with hearbeat only for 2
>daemon (freeradius & samba). I only need that the requests which
>arrive to the server are shared between the 2 servers.
>
>

Sometimes people are so used to looking for solutions to complex
problems, that they become unable to see simple solutions :-)

The point here is that you need to configure heartbeat only for the
services that need to switch machines. In casu, the only thing that
needs to switch is the ip address, and a service that distributes
requests over servers and monitors which servers have the service
running - ldirectord. Freeradius and samba should be started in the
normal way, using the system initscripts.




--
Psychics will soon lead dogs to your body.
--

Public GPG key at blackhole.pca.dfn.de

GCS/IT d- s:+ a- C(+++)$ UL++++$ P+++(++++)$ L++(+++)$ !E- W+(+++)$
N+(++) o K w$ !O !M V PS(++)@ PE-(++)@ Y+ PGP++(+++) t(+) 5 X R tv--
b++(++++) DI++(++++) D++ G e++>+++++ h(+) r y+**
Attachments: signature.asc (0.18 KB)


emi2fast at gmail

Jun 9, 2012, 4:51 AM

Post #9 of 11 (693 views)
Permalink
Re: Active/Active Cluster [In reply to]

Why you are using cman & corosync together?

I think you should use cman+pacemaker or corosync+pacemaker



2012/6/9 Yount, William D <Yount.William [at] menloworldwide>

> I have two servers which are both Dell 990's. Each server has two 1tb hard
> drives configured in RAID0. I have installed CentOS on both and they have
> the same partition sizes. I am using /dev/KNTCLFS00X/Storage as a drbd
> partition and attaching it to /dev/drbd0. DRBD syncing appears to working
> fine.
>
> I am trying to setup an Active/Active cluster. I have set up CMAN. I want
> to use this cluster just for NFS storage.
> I want to have these services running on both nodes at the same time:
>
> * IP Address
>
> * DRBD
>
> * Filesystem (gcfs2)
>
> Through a combination of official documentation and LCMC, I have this
> setup. However, I am getting this:
>
> ============
> Last updated: Fri Jun 8 23:11:58 2012
> Last change: Fri Jun 8 23:11:37 2012 via crmd on KNTCLFS002
> Stack: cman
> Current DC: KNTCLFS002 - partition with quorum
> Version: 1.1.6-3.el6-a02c0f19a00c1eb2527ad38f146ebc0834814558
> 2 Nodes configured, 2 expected votes
> 6 Resources configured.
> ============
>
> Online: [ KNTCLFS001 KNTCLFS002 ]
>
> Clone Set: cl_IPaddr2_1 [res_IPaddr2_1]
> Started: [ KNTCLFS002 KNTCLFS001 ]
> Master/Slave Set: ms_drbd_2 [res_drbd_2]
> Masters: [ KNTCLFS002 ]
> Stopped: [ res_drbd_2:0 ]
> Clone Set: cl_Filesystem_1 [res_Filesystem_1] (unique)
> res_Filesystem_1:0 (ocf::heartbeat:Filesystem): Stopped
> res_Filesystem_1:1 (ocf::heartbeat:Filesystem): Started KNTCLFS002
>
> Failed actions:
> res_Filesystem_1:0_start_0 (node=KNTCLFS002, call=42, rc=1,
> status=complete): unknown error
> res_Filesystem_1:0_start_0 (node=KNTCLFS001, call=87, rc=1,
> status=complete): unknown error
>
> I have attached my nfs.res, cluster.conf and corosync.conf files. Please
> let me know if I can provide any other information to help resolve this.
>
>
>
> Thanks,
> William Yount | Systems Analyst | Menlo Worldwide | Cell: 901-654-9933
> Safety | Leadership | Integrity | Commitment | Excellence
> Please consider the environment before printing this e-mail
>
>
> _______________________________________________
> Linux-HA mailing list
> Linux-HA [at] lists
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
>



--
esta es mi vida e me la vivo hasta que dios quiera
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems


andrew at beekhof

Jun 11, 2012, 6:53 PM

Post #10 of 11 (684 views)
Permalink
Re: Active/Active Cluster [In reply to]

On Sat, Jun 9, 2012 at 9:51 PM, emmanuel segura <emi2fast [at] gmail> wrote:
> Why you are using cman & corosync together?
>
> I think you should use cman+pacemaker or corosync+pacemaker

Right. cman uses corosync underneath, but you should only
configure+start one of them.
Probably cman in this case.

>
>
>
> 2012/6/9 Yount, William D <Yount.William [at] menloworldwide>
>
>> I have two servers which are both Dell 990's. Each server has two 1tb hard
>> drives configured in RAID0. I have installed CentOS on both and they have
>> the same partition sizes. I am using /dev/KNTCLFS00X/Storage as a drbd
>> partition and attaching it to /dev/drbd0. DRBD syncing appears to working
>> fine.
>>
>> I am trying to setup an Active/Active cluster. I have set up CMAN. I want
>> to use this cluster just for NFS storage.
>> I want to have these services running on both nodes at the same time:
>>
>> *         IP Address
>>
>> *         DRBD
>>
>> *         Filesystem (gcfs2)
>>
>> Through a combination of official documentation and LCMC, I have this
>> setup. However, I am getting this:
>>
>> ============
>> Last updated: Fri Jun  8 23:11:58 2012
>> Last change: Fri Jun  8 23:11:37 2012 via crmd on KNTCLFS002
>> Stack: cman
>> Current DC: KNTCLFS002 - partition with quorum
>> Version: 1.1.6-3.el6-a02c0f19a00c1eb2527ad38f146ebc0834814558
>> 2 Nodes configured, 2 expected votes
>> 6 Resources configured.
>> ============
>>
>> Online: [ KNTCLFS001 KNTCLFS002 ]
>>
>> Clone Set: cl_IPaddr2_1 [res_IPaddr2_1]
>>     Started: [ KNTCLFS002 KNTCLFS001 ]
>> Master/Slave Set: ms_drbd_2 [res_drbd_2]
>>     Masters: [ KNTCLFS002 ]
>>     Stopped: [ res_drbd_2:0 ]
>> Clone Set: cl_Filesystem_1 [res_Filesystem_1] (unique)
>>     res_Filesystem_1:0 (ocf::heartbeat:Filesystem):    Stopped
>>     res_Filesystem_1:1 (ocf::heartbeat:Filesystem):    Started KNTCLFS002
>>
>> Failed actions:
>>    res_Filesystem_1:0_start_0 (node=KNTCLFS002, call=42, rc=1,
>> status=complete): unknown error
>>    res_Filesystem_1:0_start_0 (node=KNTCLFS001, call=87, rc=1,
>> status=complete): unknown error
>>
>> I have attached my nfs.res, cluster.conf and corosync.conf files.  Please
>> let me know if I can provide any other information to help resolve this.
>>
>>
>>
>> Thanks,
>> William Yount | Systems Analyst | Menlo Worldwide | Cell: 901-654-9933
>> Safety | Leadership | Integrity | Commitment | Excellence
>> Please consider the environment before printing this e-mail
>>
>>
>> _______________________________________________
>> Linux-HA mailing list
>> Linux-HA [at] lists
>> http://lists.linux-ha.org/mailman/listinfo/linux-ha
>> See also: http://linux-ha.org/ReportingProblems
>>
>
>
>
> --
> esta es mi vida e me la vivo hasta que dios quiera
> _______________________________________________
> Linux-HA mailing list
> Linux-HA [at] lists
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems


andrew at beekhof

Jun 14, 2012, 9:41 PM

Post #11 of 11 (666 views)
Permalink
Re: Active/Active Cluster [In reply to]

Um, you appear to have cman and corosync as cloned resources.
Thats really not a good idea.

Have you seen the "clusters from scratch" document?
That would be a good place to start.

On Thu, Jun 14, 2012 at 8:16 AM, Yount, William D
<Yount.William [at] menloworldwide> wrote:
> I have two servers, 10.89.99.31(KNTCLFS001) and 10.89.99.32(KNTCLFS002). I am trying to use 10.89.99.30 to float between them in an Active/Active cluster.
>
>
>
> I have several services which should be running simultaneously on both servers. I am trying to set up an Active/Active cluster. I am using the cluster solely for NFS and am using GFS2 for clustered storage.
>
>
>
> What I would like is for either one of the nodes to go out, and the storage to still be available. Right now, everything works fine if KNTCLFS002 goes out. If KNTCLFS001 goes out instead, then nothing
>
> works, which probably means that everything was attaching to KNTCFLS001 to begin with. I am not sure if I should adjust my stickiness settings.
>
>
>
> One specific issue is that after a reboot, the node will start corosync but not cman. So I have to manually stop corosync and then start cman up.
>
>
>
> I am trying to use this as storage for VMs and I would like the storage to always be available. Can someone look over the attached configuration and let me know if any changes are warranted in order to achieve 'an always' up NFS share? Is there a better solution?
>
>
>
>
> Thanks,
> William
>
>
> _______________________________________________
> Linux-HA mailing list
> Linux-HA [at] lists
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Linux-HA users RSS feed   Index | Next | Previous | View Threaded
 
 


Interested in having your list archived? Contact Gossamer Threads
 
  Web Applications & Managed Hosting Powered by Gossamer Threads Inc.