[ClusterLabs] Multiple Corosync Instance on a Single Host

Steven Iveson sjiveson at live.com
Mon Jan 11 12:28:42 UTC 2016


Thank you Honza, I will give that a go and report back.

Kind regards,
Steven Iveson

> From: users-request at clusterlabs.org
> Subject: Users Digest, Vol 12, Issue 11
> To: users at clusterlabs.org
> Date: Fri, 8 Jan 2016 08:54:17 +0100
> 
> Send Users mailing list submissions to
> 	users at clusterlabs.org
> 
> To subscribe or unsubscribe via the World Wide Web, visit
> 	http://clusterlabs.org/mailman/listinfo/users
> or, via email, send a message with subject or body 'help' to
> 	users-request at clusterlabs.org
> 
> You can reach the person managing the list at
> 	users-owner at clusterlabs.org
> 
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Users digest..."
> 
> 
> Today's Topics:
> 
>    1. Re: Passing and binding to virtual IP in my service
>       (Andrei Borzenkov)
>    2. Re: Passing and binding to virtual IP in my service
>       (Andrei Borzenkov)
>    3. Multiple Corosync Instance on a Single Host (Steven Iveson)
>    4. Re: Passing and binding to virtual IP in my service
>       (Dejan Muhamedagic)
>    5. Pacemaker 1.1.14 - Release Candidate 5 (Ken Gaillot)
>    6. Parallel adding of resources (Arjun Pandey)
>    7. Re: Multiple Corosync Instance on a Single Host (Jan Friesse)
>    8. Help required for N+1 redundancy setup (Rishin Gangadharan)
> 
> 
> ----------------------------------------------------------------------
> 
> Message: 1
> Date: Thu, 7 Jan 2016 14:14:17 +0300
> From: Andrei Borzenkov <arvidjaar at gmail.com>
> To: Cluster Labs - All topics related to open-source clustering
> 	welcomed	<users at clusterlabs.org>, Nikhil Utane
> 	<nikhil.subscribed at gmail.com>
> Subject: Re: [ClusterLabs] Passing and binding to virtual IP in my
> 	service
> Message-ID: <568E4889.7000106 at gmail.com>
> Content-Type: text/plain; charset=utf-8
> 
> 07.01.2016 13:09, Kristoffer Gr?nlund ?????:
> > Nikhil Utane <nikhil.subscribed at gmail.com> writes:
> > 
> >> Of course, one way is to pass the same virtual IP configuration to my
> >> resource agent:
> >> "<nvpair id="ClusterIP-instance_attributes-ip" name="ip"
> >> value="10.206.1.253"/>"
> >> But this will be duplicate information and error-prone.
> >> There ought to be a better way. :)
> >>
> > 
> > Hi,
> > 
> > You can use attribute references to avoid having to configure the same
> > value twice. It is documented here for the crmsh syntax:
> > 
> > http://crmsh.github.io/man/#topics_Features_AttributeReferences
> > 
> 
> 
> Is it crmsh-script only or it is also stored in CIB? How would it look
> like in CIB then?
> 
> 
> 
> ------------------------------
> 
> Message: 2
> Date: Thu, 7 Jan 2016 14:19:35 +0300
> From: Andrei Borzenkov <arvidjaar at gmail.com>
> To: Cluster Labs - All topics related to open-source clustering
> 	welcomed	<users at clusterlabs.org>
> Subject: Re: [ClusterLabs] Passing and binding to virtual IP in my
> 	service
> Message-ID: <568E49C7.7090109 at gmail.com>
> Content-Type: text/plain; charset=UTF-8
> 
> 07.01.2016 13:42, Jorge F?bregas ?????:
> > On 01/07/2016 05:28 AM, Nikhil Utane wrote:
> >> So my question is, how to pass the virtual IP to my UDPSend OCF agent so
> >> that it can then bind to the vip? This will ensure that all messages
> >> initiated by my UDPSend goes from vip.
> > 
> > Hi,
> > 
> > I don't know how ping -I does it (what system call it uses) but I think
> > you'll have to implement that if you want your program to source
> > connection from a particular virtual IP.
> > 
> > As far as I know, the way this is usually done these days is by creating
> > routes.  Something like:
> > 
> > ip route change 192.168.14.0/24 dev eth0 src 192.168.14.4
> > 
> > 
> >> Out of curiosity, where is this virtual IP stored in the kernel?
> >> I expected to see a secondary interface ( for e.g. eth0:1) with the vip
> >> but it isn't there.
> > 
> > Well, in the old days we used to have a "virtual interface" (eth0:1,
> > eth0:2 etc) but the proper modern way is to use "virtual addresses"
> > within a single interface.  The caveat is that you need to use the ip
> > command to show these virtual addresses (ifconfig is not aware of them):
> 
> You can assign label to address which will then be visible in ifconfig; i.e.
> 
> ip addr add 1.2.3.4/24 label eth0:foo dev eth0
> 
> will actually make new address appear in ifconfig output as "eth0:foo"
> interface.
> 
> > 
> > ip addr show
> > 
> > You'll see there the notion of a primary IP and secondaries.  The system
> > will initiate connection from the primary by default (unless you specify
> > a route like the one above).
> > 
> > HTH,
> > Jorge
> > 
> > _______________________________________________
> > Users mailing list: Users at clusterlabs.org
> > http://clusterlabs.org/mailman/listinfo/users
> > 
> > Project Home: http://www.clusterlabs.org
> > Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> > Bugs: http://bugs.clusterlabs.org
> > 
> 
> 
> 
> 
> ------------------------------
> 
> Message: 3
> Date: Thu, 7 Jan 2016 11:00:18 +0000
> From: Steven Iveson <sjiveson at live.com>
> To: "users at clusterlabs.org" <users at clusterlabs.org>
> Subject: [ClusterLabs] Multiple Corosync Instance on a Single Host
> Message-ID: <DUB122-W4354D4C47FA212AB457DEAB2F50 at phx.gbl>
> Content-Type: text/plain; charset="iso-8859-1"
> 
> Hi list,
> I have a situation where I would like to run multiple instances of Corosync, each 'group' on a number of hosts, with each instance providing HA for a dedicated service, thus allowing for benefits around change control, risk management, service/environment separation and the like. The hosts in question run nothing but Docker containers, with the Corosync (+Pacemaker) containers using host mode networking.
> Whilst I can change the UDPU port used by Corosync, a second instance (container) fails with a socket error. Here's the relevant logs:
> Jan 06 13:47:55 [19] rancher-one corosync info    [QB    ] server name: cmapJan 06 13:47:55 [19] rancher-one corosync error   [QB    ] Could not bind AF_UNIX (): Address already in use (98)Jan 06 13:47:55 [19] rancher-one corosync info    [QB    ] withdrawing server socketsJan 06 13:47:55 [19] rancher-one corosync error   [MAIN  ] Can't initialize IPCJan 06 13:47:55 [19] rancher-one corosync error   [SERV  ] Service engine 'corosync_cmap' failed to load for reason 'qb_ipcs_run error'Jan 06 13:47:55 [19] rancher-one corosync error   [MAIN  ] Corosync Cluster Engine exiting with status 20 at service.c:356.
> Can this be overcome? Is there a way to specify a different socket?
> Many thanks in advance.
> Kind regards,
> Steven Ivesonsteve at iveson.eu@sjivesonBlogs: https://packetpushers.net/author/siveson/Books: http://www.amazon.com/Steven-Iveson/e/B00BIR04XM/07734 778 184 		 	   		  
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://clusterlabs.org/pipermail/users/attachments/20160107/b2cba53c/attachment-0001.html>
> 
> ------------------------------
> 
> Message: 4
> Date: Thu, 7 Jan 2016 17:29:25 +0100
> From: Dejan Muhamedagic <dejanmm at fastmail.fm>
> To: Cluster Labs - All topics related to open-source clustering
> 	welcomed	<users at clusterlabs.org>
> Subject: Re: [ClusterLabs] Passing and binding to virtual IP in my
> 	service
> Message-ID: <20160107162925.GA19203 at walrus.homenet>
> Content-Type: text/plain; charset=utf-8
> 
> Hi,
> 
> On Thu, Jan 07, 2016 at 02:14:17PM +0300, Andrei Borzenkov wrote:
> > 07.01.2016 13:09, Kristoffer Gr?nlund ?????:
> > > Nikhil Utane <nikhil.subscribed at gmail.com> writes:
> > > 
> > >> Of course, one way is to pass the same virtual IP configuration to my
> > >> resource agent:
> > >> "<nvpair id="ClusterIP-instance_attributes-ip" name="ip"
> > >> value="10.206.1.253"/>"
> > >> But this will be duplicate information and error-prone.
> > >> There ought to be a better way. :)
> > >>
> > > 
> > > Hi,
> > > 
> > > You can use attribute references to avoid having to configure the same
> > > value twice. It is documented here for the crmsh syntax:
> > > 
> > > http://crmsh.github.io/man/#topics_Features_AttributeReferences
> > > 
> > 
> > 
> > Is it crmsh-script only or it is also stored in CIB?
> 
> Everything ends in the CIB.
> 
> > How would it look
> > like in CIB then?
> 
> Normally, you don't need to know. I don't :)
> 
> Thanks,
> 
> Dejan
> 
> 
> > 
> > _______________________________________________
> > Users mailing list: Users at clusterlabs.org
> > http://clusterlabs.org/mailman/listinfo/users
> > 
> > Project Home: http://www.clusterlabs.org
> > Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> > Bugs: http://bugs.clusterlabs.org
> 
> 
> 
> ------------------------------
> 
> Message: 5
> Date: Thu, 7 Jan 2016 16:06:18 -0600
> From: Ken Gaillot <kgaillot at redhat.com>
> To: Cluster Labs - All topics related to open-source clustering
> 	welcomed	<users at clusterlabs.org>
> Subject: [ClusterLabs] Pacemaker 1.1.14 - Release Candidate 5
> Message-ID: <568EE15A.3010505 at redhat.com>
> Content-Type: text/plain; charset=utf-8
> 
> The source code for the latest (and possibly final) Pacemaker 1.1.14
> release candidate is available at
> https://github.com/ClusterLabs/pacemaker/releases/tag/Pacemaker-1.1.14-rc5
> 
> This release fixes a significant regression introduced in commit
> 12ca8bb: If a node had a fencing topology with multiple levels, only the
> last level would actually be used.
> 
> This release also fixes misspellings in various log messages and updates
> the SNMP MIB.
> 
> Everyone is encouraged to download, compile and test the new release.
> Your feedback is important and appreciated.
> 
> If there are no significant issues found, this will likely become the
> final 1.1.14 release next week.
> 
> P.S. The Pacemaker-1.1.14-rc4 tag was applied to the master branch by
> mistake, so please do not use it. :)
> -- 
> Ken Gaillot <kgaillot at redhat.com>
> 
> 
> 
> ------------------------------
> 
> Message: 6
> Date: Fri, 8 Jan 2016 12:04:01 +0530
> From: Arjun Pandey <apandepublic at gmail.com>
> To: Cluster Labs - All topics related to open-source clustering
> 	welcomed	<users at clusterlabs.org>
> Subject: [ClusterLabs] Parallel adding of resources
> Message-ID:
> 	<CAKpWUY=fftNfj-AJSDn7SLfS+-GAH_baJ74zK-0vQ_yqTd+S3g at mail.gmail.com>
> Content-Type: text/plain; charset=UTF-8
> 
> Hi
> 
> I am running a 2 node cluster with this config on centos 6.6
> 
> Master/Slave Set: foo-master [foo]
> Masters: [ messi ]
> Stopped: [ronaldo ]
>  eth1-CP        (ocf::pw:IPaddr):       Started messi
>  eth2-UP        (ocf::pw:IPaddr):       Started messi
>  eth3-UPCP      (ocf::pw:IPaddr):       Started messi
> 
> where i have a multi-state resource foo being run in master/slave mode
> and  IPaddr RA is just modified IPAddr2 RA. Additionally i have a
> collocation constraint for the IP addr to be collocated with the master.
> 
> Now there are cases where i have multiple virtual IP's ( around 20 )
> and for failover time gets substantially increased in these cases.
> Based on the logs what i have observed is the IPaddr resources are
> moved sequentially. Is this really the case ? Also is it possible to
> specify that they can be added simultaneously, since none of them have
> any sort of corelation with the other ?
> 
> If it's sequential what is the reason behind it ?
> 
> 
> Thanks in advance.
> 
> Regards
> Arjun
> 
> 
> 
> ------------------------------
> 
> Message: 7
> Date: Fri, 8 Jan 2016 08:47:50 +0100
> From: Jan Friesse <jfriesse at redhat.com>
> To: Cluster Labs - All topics related to open-source clustering
> 	welcomed	<users at clusterlabs.org>
> Subject: Re: [ClusterLabs] Multiple Corosync Instance on a Single Host
> Message-ID: <568F69A6.3040609 at redhat.com>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
> 
> Steven,
> 
>  > Hi list,
> > I have a situation where I would like to run multiple instances of Corosync, each 'group' on a number of hosts, with each instance providing HA for a dedicated service, thus allowing for benefits around change control, risk management, service/environment separation and the like. The hosts in question run nothing but Docker containers, with the Corosync (+Pacemaker) containers using host mode networking.
> > Whilst I can change the UDPU port used by Corosync, a second instance (container) fails with a socket error. Here's the relevant logs:
> > Jan 06 13:47:55 [19] rancher-one corosync info    [QB    ] server name: cmapJan 06 13:47:55 [19] rancher-one corosync error   [QB    ] Could not bind AF_UNIX (): Address already in use (98)Jan 06 13:47:55 [19] rancher-one corosync info    [QB    ] withdrawing server socketsJan 06 13:47:55 [19] rancher-one corosync error   [MAIN  ] Can't initialize IPCJan 06 13:47:55 [19] rancher-one corosync error   [SERV  ] Service engine 'corosync_cmap' failed to load for reason 'qb_ipcs_run error'Jan 06 13:47:55 [19] rancher-one corosync error   [MAIN  ] Corosync Cluster Engine exiting with status 20 at service.c:356.
> > Can this be overcome? Is there a way to specify a different socket?
> 
> Corosync is not designed for running multiple instances. But using 
> docker is different story (it's more or less VM, so there should be no 
> problem).
> 
> You can try to look what socket is already in use (for example use 
> strace) but I believe main problem is shared /dev/shm. Can you try to 
> mount fresh tmpfs to /dev/shm in docker container?
> 
> Regards,
>    Honza
> 
> > Many thanks in advance.
> > Kind regards,
> > Steven Ivesonsteve at iveson.eu@sjivesonBlogs: https://packetpushers.net/author/siveson/Books: http://www.amazon.com/Steven-Iveson/e/B00BIR04XM/07734 778 184 		 	   		
> >
> >
> >
> > _______________________________________________
> > Users mailing list: Users at clusterlabs.org
> > http://clusterlabs.org/mailman/listinfo/users
> >
> > Project Home: http://www.clusterlabs.org
> > Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> > Bugs: http://bugs.clusterlabs.org
> >
> 
> 
> 
> 
> ------------------------------
> 
> Message: 8
> Date: Fri, 8 Jan 2016 07:52:52 +0000
> From: Rishin Gangadharan <RG00423341 at TechMahindra.com>
> To: "'users at clusterlabs.org'" <users at clusterlabs.org>
> Subject: [ClusterLabs] Help required for N+1 redundancy setup
> Message-ID:
> 	<596f7e307e23463f8b75d8dcd0af1087 at BLREXCHMBX003.TechMahindra.com>
> Content-Type: text/plain; charset="us-ascii"
> 
> Hi
>    Can anybody tell be ,how to configure corosync pacemaker with crmsh  for active -active and N+1redundancy setup for kamailio.
> 
> Thanks
> Rishin
> 
> 
> 
> 
> 
> ============================================================================================================================
> Disclaimer:  This message and the information contained herein is proprietary and confidential and subject to the Tech Mahindra policy statement, you may review the policy at http://www.techmahindra.com/Disclaimer.html externally http://tim.techmahindra.com/tim/disclaimer.html internally within TechMahindra.
> ============================================================================================================================
> 
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <http://clusterlabs.org/pipermail/users/attachments/20160108/ed6f7dc9/attachment.html>
> 
> ------------------------------
> 
> _______________________________________________
> Users mailing list
> Users at clusterlabs.org
> http://clusterlabs.org/mailman/listinfo/users
> 
> 
> End of Users Digest, Vol 12, Issue 11
> *************************************
 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clusterlabs.org/pipermail/users/attachments/20160111/c6407db3/attachment.htm>


More information about the Users mailing list