Re: [Pacemaker] Linux HA setup for CentOS 6.5

2014-10-15 Thread Digimer 

On 16/10/14 12:31 AM, Alex Samad - Yieldbroker wrote:

I believe there was a change 6.4 -> 6.5 that moved from openais to
cman…  I believe in rhel7 it is going back..

I ran into this problem when I upgraded my centos 6.2 -> 6.5..

A


To elaborate a little on Andrew's post;

The cman package used openais in RHEL 5 for cluster communications and 
membership. In RHEL 6, openais was replaced by corosync, which was 
basically a stripped down and HA focused openais (openais itself became 
an optional corosync plugin for people who wanted the full AIS API).


RHEL7 is not "going back" at all. In fact, cman (and rgmanager) are gone 
entirely and the stack has unified on corosync + pacemaker on all major 
distros now.


I've got an incomplete history of all this here:

https://alteeve.ca/w/History_of_HA_Clustering

Cheers

--
Digimer
Papers and Projects: https://alteeve.ca/w/
What if the cure for cancer is trapped in the mind of a person without 
access to education?


___
Pacemaker mailing list: Pacemaker@oss.clusterlabs.org
http://oss.clusterlabs.org/mailman/listinfo/pacemaker

Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://bugs.clusterlabs.org

Re: [Pacemaker] Linux HA setup for CentOS 6.5

2014-10-15 Thread Andrew Beekhof 

On 16 Oct 2014, at 3:31 pm, Alex Samad - Yieldbroker 
 wrote:

> I believe there was a change 6.4 -> 6.5 that moved from openais to cman…  I 
> believe in rhel7 it is going back..

No, its going to native corosync (which is quite different to running against a 
custom plugin loaded into openais).

'cman' here is pretty much openais with a different config file and init script

>  
> I ran into this problem when I upgraded my centos 6.2 -> 6.5..
>  
> A
>  
> From: Sihan Goi [mailto:gois...@gmail.com] 
> Sent: Thursday, 16 October 2014 3:15 PM
> To: The Pacemaker cluster resource manager
> Subject: Re: [Pacemaker] Linux HA setup for CentOS 6.5
>  
> After following the guide, I've successfully managed to get Apache server up 
> and running in the cluster as an active/passive setup, but with some 
> differences. My cluster stack is stated as being cman while the guide's is 
> openais. Not sure if that's a problem. Also, some commands in the guide don't 
> seem to work.
> 
> I'm moving on to DRBD installation now, but when I do a "yum install 
> drbd-pacemaker drbd-udev", these packages are not available. After some 
> googling, it seems that drbd83-utils/kmod-drbd83 or drbd84-utils/kmod-drbd84 
> is available via another repo. Does this work with the guide?
>  
> On Thu, Oct 16, 2014 at 9:35 AM, Sihan Goi  wrote:
> Hi,
> 
> Thanks for the guide! I thought I had the same exact version...mine is also 
> named "Pacemaker 1.1 Clusters from Scratch Creating Active/Passive and 
> Active/Active Clusters on Fedora Edition 5", but my version of the document 
> is meant for Fedora 17, and uses pcs and systemctl calls which don't exist on 
> CentOS 6.5. I was trying to get it to work on CentOS 7 but realized support 
> for DRBD on CentOS 7 is really lacking.
> 
> I'll refer to the version you posted from hereon.
>  
> On Wed, Oct 15, 2014 at 11:43 PM, Digimer  wrote:
> Let pacemaker start cman/corosync on EL6.
> 
> This is the guide that covers it, written by Pacemaker's author:
> 
> http://clusterlabs.org/doc/en-US/Pacemaker/1.1-plugin/html-single/Clusters_from_Scratch/index.html
> 
> It notes that it's based on Fedora 13, but that maps to EL6 almost perfectly.
> 
> A very slightly altered approach is here, in my *very* unfinished tutorial:
> 
> https://alteeve.ca/w/Anvil!_Tutorial_3_on_EL6#Configuring_the_Anvil.21
> 
> The main difference is that Andrew's approach (see section 8.2.2) is to 
> disable quorum via editing /etc/sysconfig/cman, where my approach handles it 
> in the main /etc/cluster/cluster.conf (cman's main config file).
> 
> In any case, from then on, start pacemaker and let it handle everything else.
> 
> Cheers
> 
> digimer
> 
> On 15/10/14 04:27 AM, Sihan Goi wrote:
> Hi,
> 
> So I've decided to make things simpler and go with a wired network
> instead of wireless. I connected both boxes to a router, manually edited
> the ifcfg-eth0 files to set static IP addresses for both boxes (not
> before downloading and building a driver for the nic of 1 of the boxes),
> did a "chkconfig NetworkManager off", "service NetworkManager stop", and
> "service network restart".
> 
> I'm able to ping each other via IP address and hostname. I also already
> have corosync, pacemaker, crmsh and cman installed.
> 
> I then did the following as per the guide at
> http://geekpeek.net/linux-cluster-corosync-pacemaker
> 
> service corosync start - success.
> service pacemaker start - I get a "Starting cman...corosync cluster
> engine is already running [FAILED]"
> 
> What's up? :(
> 
> On Oct 15, 2014 12:23 PM, "Sihan Goi"  > wrote:
> 
> No typo.
> 
> [root@node02 network-scripts]# ls -lah
> /etc/sysconfig/network-scripts/ifcfg-*
> -rw-r--r--. 1 root root 254 Oct 10  2013
> /etc/sysconfig/network-scripts/ifcfg-lo
> 
> I installed CentOS 6.5 with the LiveDVD. I found it weird as well
> that these files were missing.
> 
> On Wed, Oct 15, 2014 at 11:54 AM, Digimer  > wrote:
> 
> Sure there isn't a typo there?
> 
> an-c05n01:~# ls -lah /etc/sysconfig/network-__scripts/ifcfg-*
> -rw-r--r--. 1 root root 225 Jan 16  2013
> /etc/sysconfig/network-__scripts/ifcfg-bond0
> -rw-r--r--. 1 root root 220 Jan 16  2013
> /etc/sysconfig/network-__scripts/ifcfg-bond1
> -rw-r--r--. 1 root root 198 Jan 16  2013
> /etc/sysconfig/network-__scripts/ifcfg-bond2
> -rw-r--r--. 1 root root 149 Jan 16  2013
> /etc/sysconfig/network-__scripts/ifcfg-eth0
> -rw-r--r--. 1 root root 144 Jan 16  2013
> /etc/sysconfig/network-__scripts/ifcfg-eth1
> -rw-r--r--. 1 root root 152 Mar 14  2013
> /etc/sysconfig/network-__scripts/ifcfg-eth2
> -rw-r--r--. 1 root root 149 Jan 16  2013
> /etc/sysconfig/network-__scripts/ifcfg-eth3
> -rw-r--r--. 1 root root 144 Jan 16  2013
> /etc/sysconfig/network-__scripts/ifcfg-eth4
> -rw-r--r--. 1 r

Re: [Pacemaker] Linux HA setup for CentOS 6.5

2014-10-15 Thread Digimer 

On 16/10/14 12:14 AM, Sihan Goi wrote:

After following the guide, I've successfully managed to get Apache
server up and running in the cluster as an active/passive setup, but
with some differences. My cluster stack is stated as being cman while
the guide's is openais. Not sure if that's a problem. Also, some
commands in the guide don't seem to work.


If you can provide examples of what issues you're having, I will be 
happy to try an help.



I'm moving on to DRBD installation now, but when I do a "yum install
drbd-pacemaker drbd-udev", these packages are not available. After some
googling, it seems that drbd83-utils/kmod-drbd83 or
drbd84-utils/kmod-drbd84 is available via another repo. Does this work
with the guide?


You need to get them from a 3rd party repo (or install from source). I 
personally still use 8.3.16 (consistency during "Anvil!" generations), 
but I know that 8.4 is fine on EL6 (and EL7, to address an earlier 
comment). I have my own repos with these packages, but you would likely 
be better served using the ELRepo ones.


https://alteeve.ca/w/AN!Cluster_Tutorial_2#Installing_DRBD

The only real difference is to s/83/84/:

+ yum install drbd84-utils kmod-drbd84
- yum install drbd83-utils kmod-drbd83

If you run into any troubles, please share details and I am sure we'll 
get you sorted out in no time.


Cheers

--
Digimer
Papers and Projects: https://alteeve.ca/w/
What if the cure for cancer is trapped in the mind of a person without 
access to education?


___
Pacemaker mailing list: Pacemaker@oss.clusterlabs.org
http://oss.clusterlabs.org/mailman/listinfo/pacemaker

Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://bugs.clusterlabs.org

Re: [Pacemaker] Linux HA setup for CentOS 6.5

2014-10-15 Thread Alex Samad - Yieldbroker 
I believe there was a change 6.4 -> 6.5 that moved from openais to cman…  I 
believe in rhel7 it is going back..

I ran into this problem when I upgraded my centos 6.2 -> 6.5..

A

From: Sihan Goi [mailto:gois...@gmail.com]
Sent: Thursday, 16 October 2014 3:15 PM
To: The Pacemaker cluster resource manager
Subject: Re: [Pacemaker] Linux HA setup for CentOS 6.5

After following the guide, I've successfully managed to get Apache server up 
and running in the cluster as an active/passive setup, but with some 
differences. My cluster stack is stated as being cman while the guide's is 
openais. Not sure if that's a problem. Also, some commands in the guide don't 
seem to work.

I'm moving on to DRBD installation now, but when I do a "yum install 
drbd-pacemaker drbd-udev", these packages are not available. After some 
googling, it seems that drbd83-utils/kmod-drbd83 or drbd84-utils/kmod-drbd84 is 
available via another repo. Does this work with the guide?

On Thu, Oct 16, 2014 at 9:35 AM, Sihan Goi 
mailto:gois...@gmail.com>> wrote:
Hi,

Thanks for the guide! I thought I had the same exact version...mine is also 
named "Pacemaker 1.1 Clusters from Scratch Creating Active/Passive and 
Active/Active Clusters on Fedora Edition 5", but my version of the document is 
meant for Fedora 17, and uses pcs and systemctl calls which don't exist on 
CentOS 6.5. I was trying to get it to work on CentOS 7 but realized support for 
DRBD on CentOS 7 is really lacking.

I'll refer to the version you posted from hereon.

On Wed, Oct 15, 2014 at 11:43 PM, Digimer 
mailto:li...@alteeve.ca>> wrote:
Let pacemaker start cman/corosync on EL6.

This is the guide that covers it, written by Pacemaker's author:

http://clusterlabs.org/doc/en-US/Pacemaker/1.1-plugin/html-single/Clusters_from_Scratch/index.html

It notes that it's based on Fedora 13, but that maps to EL6 almost perfectly.

A very slightly altered approach is here, in my *very* unfinished tutorial:

https://alteeve.ca/w/Anvil!_Tutorial_3_on_EL6#Configuring_the_Anvil.21

The main difference is that Andrew's approach (see section 8.2.2) is to disable 
quorum via editing /etc/sysconfig/cman, where my approach handles it in the 
main /etc/cluster/cluster.conf (cman's main config file).

In any case, from then on, start pacemaker and let it handle everything else.

Cheers

digimer

On 15/10/14 04:27 AM, Sihan Goi wrote:
Hi,

So I've decided to make things simpler and go with a wired network
instead of wireless. I connected both boxes to a router, manually edited
the ifcfg-eth0 files to set static IP addresses for both boxes (not
before downloading and building a driver for the nic of 1 of the boxes),
did a "chkconfig NetworkManager off", "service NetworkManager stop", and
"service network restart".

I'm able to ping each other via IP address and hostname. I also already
have corosync, pacemaker, crmsh and cman installed.

I then did the following as per the guide at
http://geekpeek.net/linux-cluster-corosync-pacemaker

service corosync start - success.
service pacemaker start - I get a "Starting cman...corosync cluster
engine is already running [FAILED]"

What's up? :(

On Oct 15, 2014 12:23 PM, "Sihan Goi" 
mailto:gois...@gmail.com>
>> wrote:

No typo.

[root@node02 network-scripts]# ls -lah
/etc/sysconfig/network-scripts/ifcfg-*
-rw-r--r--. 1 root root 254 Oct 10  2013
/etc/sysconfig/network-scripts/ifcfg-lo

I installed CentOS 6.5 with the LiveDVD. I found it weird as well
that these files were missing.

On Wed, Oct 15, 2014 at 11:54 AM, Digimer 
mailto:li...@alteeve.ca>
>> wrote:

Sure there isn't a typo there?

an-c05n01:~# ls -lah /etc/sysconfig/network-__scripts/ifcfg-*
-rw-r--r--. 1 root root 225 Jan 16  2013
/etc/sysconfig/network-__scripts/ifcfg-bond0
-rw-r--r--. 1 root root 220 Jan 16  2013
/etc/sysconfig/network-__scripts/ifcfg-bond1
-rw-r--r--. 1 root root 198 Jan 16  2013
/etc/sysconfig/network-__scripts/ifcfg-bond2
-rw-r--r--. 1 root root 149 Jan 16  2013
/etc/sysconfig/network-__scripts/ifcfg-eth0
-rw-r--r--. 1 root root 144 Jan 16  2013
/etc/sysconfig/network-__scripts/ifcfg-eth1
-rw-r--r--. 1 root root 152 Mar 14  2013
/etc/sysconfig/network-__scripts/ifcfg-eth2
-rw-r--r--. 1 root root 149 Jan 16  2013
/etc/sysconfig/network-__scripts/ifcfg-eth3
-rw-r--r--. 1 root root 144 Jan 16  2013
/etc/sysconfig/network-__scripts/ifcfg-eth4
-rw-r--r--. 1 root root 152 Mar 14  2013
/etc/sysconfig/network-__scripts/ifcfg-eth5
-rw-r--r--. 1 root root 254 Jul 22 09:56
/etc/sysconfig/network-__scripts/ifcfg-lo
-rw-r--r--. 1 root root 213 Mar 13  2013
/etc/sysconfig/network-__scripts/ifcfg-vbr2

I've never seen an EL6 install without the files there,

Re: [Pacemaker] Linux HA setup for CentOS 6.5

2014-10-15 Thread Sihan Goi 
After following the guide, I've successfully managed to get Apache server
up and running in the cluster as an active/passive setup, but with some
differences. My cluster stack is stated as being cman while the guide's is
openais. Not sure if that's a problem. Also, some commands in the guide
don't seem to work.

I'm moving on to DRBD installation now, but when I do a "yum install
drbd-pacemaker drbd-udev", these packages are not available. After some
googling, it seems that drbd83-utils/kmod-drbd83 or
drbd84-utils/kmod-drbd84 is available via another repo. Does this work with
the guide?

On Thu, Oct 16, 2014 at 9:35 AM, Sihan Goi  wrote:

> Hi,
>
> Thanks for the guide! I thought I had the same exact version...mine is
> also named "Pacemaker 1.1 Clusters from Scratch Creating Active/Passive and
> Active/Active Clusters on Fedora Edition 5", but my version of the document
> is meant for Fedora 17, and uses pcs and systemctl calls which don't exist
> on CentOS 6.5. I was trying to get it to work on CentOS 7 but realized
> support for DRBD on CentOS 7 is really lacking.
>
> I'll refer to the version you posted from hereon.
>
> On Wed, Oct 15, 2014 at 11:43 PM, Digimer  wrote:
>
>> Let pacemaker start cman/corosync on EL6.
>>
>> This is the guide that covers it, written by Pacemaker's author:
>>
>> http://clusterlabs.org/doc/en-US/Pacemaker/1.1-plugin/html-
>> single/Clusters_from_Scratch/index.html
>>
>> It notes that it's based on Fedora 13, but that maps to EL6 almost
>> perfectly.
>>
>> A very slightly altered approach is here, in my *very* unfinished
>> tutorial:
>>
>> https://alteeve.ca/w/Anvil!_Tutorial_3_on_EL6#Configuring_the_Anvil.21
>>
>> The main difference is that Andrew's approach (see section 8.2.2) is to
>> disable quorum via editing /etc/sysconfig/cman, where my approach handles
>> it in the main /etc/cluster/cluster.conf (cman's main config file).
>>
>> In any case, from then on, start pacemaker and let it handle everything
>> else.
>>
>> Cheers
>>
>> digimer
>>
>> On 15/10/14 04:27 AM, Sihan Goi wrote:
>>
>>> Hi,
>>>
>>> So I've decided to make things simpler and go with a wired network
>>> instead of wireless. I connected both boxes to a router, manually edited
>>> the ifcfg-eth0 files to set static IP addresses for both boxes (not
>>> before downloading and building a driver for the nic of 1 of the boxes),
>>> did a "chkconfig NetworkManager off", "service NetworkManager stop", and
>>> "service network restart".
>>>
>>> I'm able to ping each other via IP address and hostname. I also already
>>> have corosync, pacemaker, crmsh and cman installed.
>>>
>>> I then did the following as per the guide at
>>> http://geekpeek.net/linux-cluster-corosync-pacemaker
>>>
>>> service corosync start - success.
>>> service pacemaker start - I get a "Starting cman...corosync cluster
>>> engine is already running [FAILED]"
>>>
>>> What's up? :(
>>>
>>> On Oct 15, 2014 12:23 PM, "Sihan Goi" >> > wrote:
>>>
>>> No typo.
>>>
>>> [root@node02 network-scripts]# ls -lah
>>> /etc/sysconfig/network-scripts/ifcfg-*
>>> -rw-r--r--. 1 root root 254 Oct 10  2013
>>> /etc/sysconfig/network-scripts/ifcfg-lo
>>>
>>> I installed CentOS 6.5 with the LiveDVD. I found it weird as well
>>> that these files were missing.
>>>
>>> On Wed, Oct 15, 2014 at 11:54 AM, Digimer >> > wrote:
>>>
>>> Sure there isn't a typo there?
>>>
>>> an-c05n01:~# ls -lah /etc/sysconfig/network-__scripts/ifcfg-*
>>> -rw-r--r--. 1 root root 225 Jan 16  2013
>>> /etc/sysconfig/network-__scripts/ifcfg-bond0
>>> -rw-r--r--. 1 root root 220 Jan 16  2013
>>> /etc/sysconfig/network-__scripts/ifcfg-bond1
>>> -rw-r--r--. 1 root root 198 Jan 16  2013
>>> /etc/sysconfig/network-__scripts/ifcfg-bond2
>>> -rw-r--r--. 1 root root 149 Jan 16  2013
>>> /etc/sysconfig/network-__scripts/ifcfg-eth0
>>> -rw-r--r--. 1 root root 144 Jan 16  2013
>>> /etc/sysconfig/network-__scripts/ifcfg-eth1
>>> -rw-r--r--. 1 root root 152 Mar 14  2013
>>> /etc/sysconfig/network-__scripts/ifcfg-eth2
>>> -rw-r--r--. 1 root root 149 Jan 16  2013
>>> /etc/sysconfig/network-__scripts/ifcfg-eth3
>>> -rw-r--r--. 1 root root 144 Jan 16  2013
>>> /etc/sysconfig/network-__scripts/ifcfg-eth4
>>> -rw-r--r--. 1 root root 152 Mar 14  2013
>>> /etc/sysconfig/network-__scripts/ifcfg-eth5
>>> -rw-r--r--. 1 root root 254 Jul 22 09:56
>>> /etc/sysconfig/network-__scripts/ifcfg-lo
>>> -rw-r--r--. 1 root root 213 Mar 13  2013
>>> /etc/sysconfig/network-__scripts/ifcfg-vbr2
>>>
>>> I've never seen an EL6 install without the files there,
>>> 'network' or NetworkManager aside.
>>>
>>> digimer
>>>
>>> On 14/10/14 11:32 PM, Sihan Goi wrote:
>>>
>>> There aren't any config files in
>>>

Re: [Pacemaker] Linux HA setup for CentOS 6.5

2014-10-15 Thread Sihan Goi 
Hi,

Thanks for the guide! I thought I had the same exact version...mine is also
named "Pacemaker 1.1 Clusters from Scratch Creating Active/Passive and
Active/Active Clusters on Fedora Edition 5", but my version of the document
is meant for Fedora 17, and uses pcs and systemctl calls which don't exist
on CentOS 6.5. I was trying to get it to work on CentOS 7 but realized
support for DRBD on CentOS 7 is really lacking.

I'll refer to the version you posted from hereon.

On Wed, Oct 15, 2014 at 11:43 PM, Digimer  wrote:

> Let pacemaker start cman/corosync on EL6.
>
> This is the guide that covers it, written by Pacemaker's author:
>
> http://clusterlabs.org/doc/en-US/Pacemaker/1.1-plugin/html-
> single/Clusters_from_Scratch/index.html
>
> It notes that it's based on Fedora 13, but that maps to EL6 almost
> perfectly.
>
> A very slightly altered approach is here, in my *very* unfinished tutorial:
>
> https://alteeve.ca/w/Anvil!_Tutorial_3_on_EL6#Configuring_the_Anvil.21
>
> The main difference is that Andrew's approach (see section 8.2.2) is to
> disable quorum via editing /etc/sysconfig/cman, where my approach handles
> it in the main /etc/cluster/cluster.conf (cman's main config file).
>
> In any case, from then on, start pacemaker and let it handle everything
> else.
>
> Cheers
>
> digimer
>
> On 15/10/14 04:27 AM, Sihan Goi wrote:
>
>> Hi,
>>
>> So I've decided to make things simpler and go with a wired network
>> instead of wireless. I connected both boxes to a router, manually edited
>> the ifcfg-eth0 files to set static IP addresses for both boxes (not
>> before downloading and building a driver for the nic of 1 of the boxes),
>> did a "chkconfig NetworkManager off", "service NetworkManager stop", and
>> "service network restart".
>>
>> I'm able to ping each other via IP address and hostname. I also already
>> have corosync, pacemaker, crmsh and cman installed.
>>
>> I then did the following as per the guide at
>> http://geekpeek.net/linux-cluster-corosync-pacemaker
>>
>> service corosync start - success.
>> service pacemaker start - I get a "Starting cman...corosync cluster
>> engine is already running [FAILED]"
>>
>> What's up? :(
>>
>> On Oct 15, 2014 12:23 PM, "Sihan Goi" > > wrote:
>>
>> No typo.
>>
>> [root@node02 network-scripts]# ls -lah
>> /etc/sysconfig/network-scripts/ifcfg-*
>> -rw-r--r--. 1 root root 254 Oct 10  2013
>> /etc/sysconfig/network-scripts/ifcfg-lo
>>
>> I installed CentOS 6.5 with the LiveDVD. I found it weird as well
>> that these files were missing.
>>
>> On Wed, Oct 15, 2014 at 11:54 AM, Digimer > > wrote:
>>
>> Sure there isn't a typo there?
>>
>> an-c05n01:~# ls -lah /etc/sysconfig/network-__scripts/ifcfg-*
>> -rw-r--r--. 1 root root 225 Jan 16  2013
>> /etc/sysconfig/network-__scripts/ifcfg-bond0
>> -rw-r--r--. 1 root root 220 Jan 16  2013
>> /etc/sysconfig/network-__scripts/ifcfg-bond1
>> -rw-r--r--. 1 root root 198 Jan 16  2013
>> /etc/sysconfig/network-__scripts/ifcfg-bond2
>> -rw-r--r--. 1 root root 149 Jan 16  2013
>> /etc/sysconfig/network-__scripts/ifcfg-eth0
>> -rw-r--r--. 1 root root 144 Jan 16  2013
>> /etc/sysconfig/network-__scripts/ifcfg-eth1
>> -rw-r--r--. 1 root root 152 Mar 14  2013
>> /etc/sysconfig/network-__scripts/ifcfg-eth2
>> -rw-r--r--. 1 root root 149 Jan 16  2013
>> /etc/sysconfig/network-__scripts/ifcfg-eth3
>> -rw-r--r--. 1 root root 144 Jan 16  2013
>> /etc/sysconfig/network-__scripts/ifcfg-eth4
>> -rw-r--r--. 1 root root 152 Mar 14  2013
>> /etc/sysconfig/network-__scripts/ifcfg-eth5
>> -rw-r--r--. 1 root root 254 Jul 22 09:56
>> /etc/sysconfig/network-__scripts/ifcfg-lo
>> -rw-r--r--. 1 root root 213 Mar 13  2013
>> /etc/sysconfig/network-__scripts/ifcfg-vbr2
>>
>> I've never seen an EL6 install without the files there,
>> 'network' or NetworkManager aside.
>>
>> digimer
>>
>> On 14/10/14 11:32 PM, Sihan Goi wrote:
>>
>> There aren't any config files in
>> /etc/sysconfig/network-__scripts. When I
>> was using CentOS 7, the config files were there
>> (ifcfg-something) but in
>> this CentOS 6.5 installation, they are missing.
>>
>> If is possible to not use cman, and just use corosync and
>> pacemaker? If
>> so, how?
>>
>> On Wed, Oct 15, 2014 at 11:22 AM, Digimer > 
>> >> wrote:
>>
>>  You can manually configure the wireless LAN without
>> NetworkManager.
>>  If you take a look, there should be existing config
>> files in
>>  /etc/sysconfig/network-scripts/ fo

Re: [Pacemaker] Linux HA setup for CentOS 6.5

2014-10-15 Thread Andrew Beekhof 

On 15 Oct 2014, at 7:27 pm, Sihan Goi  wrote:

> Hi,
> 
> So I've decided to make things simpler and go with a wired network instead of 
> wireless. I connected both boxes to a router, manually edited the ifcfg-eth0 
> files to set static IP addresses for both boxes (not before downloading and 
> building a driver for the nic of 1 of the boxes), did a "chkconfig 
> NetworkManager off", "service NetworkManager stop", and "service network 
> restart".
> 
> I'm able to ping each other via IP address and hostname. I also already have 
> corosync, pacemaker, crmsh and cman installed.
> 
> I then did the following as per the guide at 
> http://geekpeek.net/linux-cluster-corosync-pacemaker
> 
> service corosync start - success.

don't do this.
either just 'service pacemaker start' or, if you must, run 'service cman start' 
instead.

cman != corosync

> service pacemaker start - I get a "Starting cman...corosync cluster engine is 
> already running [FAILED]"
> 
> What's up? :(
> 
> On Oct 15, 2014 12:23 PM, "Sihan Goi"  wrote:
> No typo.
> 
> [root@node02 network-scripts]# ls -lah /etc/sysconfig/network-scripts/ifcfg-*
> -rw-r--r--. 1 root root 254 Oct 10  2013 
> /etc/sysconfig/network-scripts/ifcfg-lo
> 
> I installed CentOS 6.5 with the LiveDVD. I found it weird as well that these 
> files were missing.
> 
> On Wed, Oct 15, 2014 at 11:54 AM, Digimer  wrote:
> Sure there isn't a typo there?
> 
> an-c05n01:~# ls -lah /etc/sysconfig/network-scripts/ifcfg-*
> -rw-r--r--. 1 root root 225 Jan 16  2013 
> /etc/sysconfig/network-scripts/ifcfg-bond0
> -rw-r--r--. 1 root root 220 Jan 16  2013 
> /etc/sysconfig/network-scripts/ifcfg-bond1
> -rw-r--r--. 1 root root 198 Jan 16  2013 
> /etc/sysconfig/network-scripts/ifcfg-bond2
> -rw-r--r--. 1 root root 149 Jan 16  2013 
> /etc/sysconfig/network-scripts/ifcfg-eth0
> -rw-r--r--. 1 root root 144 Jan 16  2013 
> /etc/sysconfig/network-scripts/ifcfg-eth1
> -rw-r--r--. 1 root root 152 Mar 14  2013 
> /etc/sysconfig/network-scripts/ifcfg-eth2
> -rw-r--r--. 1 root root 149 Jan 16  2013 
> /etc/sysconfig/network-scripts/ifcfg-eth3
> -rw-r--r--. 1 root root 144 Jan 16  2013 
> /etc/sysconfig/network-scripts/ifcfg-eth4
> -rw-r--r--. 1 root root 152 Mar 14  2013 
> /etc/sysconfig/network-scripts/ifcfg-eth5
> -rw-r--r--. 1 root root 254 Jul 22 09:56 
> /etc/sysconfig/network-scripts/ifcfg-lo
> -rw-r--r--. 1 root root 213 Mar 13  2013 
> /etc/sysconfig/network-scripts/ifcfg-vbr2
> 
> I've never seen an EL6 install without the files there, 'network' or 
> NetworkManager aside.
> 
> digimer
> 
> On 14/10/14 11:32 PM, Sihan Goi wrote:
> There aren't any config files in /etc/sysconfig/network-scripts. When I
> was using CentOS 7, the config files were there (ifcfg-something) but in
> this CentOS 6.5 installation, they are missing.
> 
> If is possible to not use cman, and just use corosync and pacemaker? If
> so, how?
> 
> On Wed, Oct 15, 2014 at 11:22 AM, Digimer  > wrote:
> 
> You can manually configure the wireless LAN without NetworkManager.
> If you take a look, there should be existing config files in
> /etc/sysconfig/network-__scripts/ for the wireless connection. I've
> not done it myself since many Fedora's ago, but I believe you can
> change NMCONTROLLER="no" and then start it up with
> /etc/sysconfig/network start. I could be a bit wrong, but I am sure
> you can make wireless work without NM.
> 
> Question; Servers with WLAN? I assume these won't be used for corosync?
> 
> digimer
> 
> 
> On 14/10/14 11:17 PM, Sihan Goi wrote:
> 
> Hi,
> 
> Is there a tutorial showing how to get a basic Linux HA setup with
> replicated storage (via DRBD) working on CentOS 6.5? I want to have
> mySQL as the HA resource with the database replicated across the
> nodes.
> I've scoured the web for one but it seems that I get stuck in
> each one
> somewhere.
> 
> To elaborate, I have 2 CentOS 6.5 nodes configured with distinct
> hostnames and static IPs. They are connected to a wireless AP,
> and can
> ping each other.
> 
> I tried following this guide -
> http://clusterlabs.org/__quickstart-redhat.html
> 
> However, cman will not start when NetworkManager is running, and my
> nodes cannot connect to the wireless AP without NetworkManager
> running.
> Am I missing something or is that the stupidest dependency ever?
> How is
> a cluster supposed to work when the nodes aren't connected to
> one another?
> 
> I also tried following the "clusters from scratch" guide but
> that seems
> to rely on systemctl calls which aren't available on CentOS 6.5.
> 
> Any help?
> 
> --
> - Goi Sihan
> gois...@gmail.com 
>

Re: [Pacemaker] Pacemaker Corosync Issue

2014-10-15 Thread Andrew Beekhof 
You don't think that might be a little short?
Any failure that happened more than 10s is going to be ignored, leading to the 
pengine message you saw.

On 16 Oct 2014, at 12:21 am, Sahil Aggarwal  wrote:

> failure timeout for resource is 10s. 
> 
> On Wed, Oct 15, 2014 at 2:51 AM, Andrew Beekhof  wrote:
> 
> On 15 Oct 2014, at 4:23 am, Sahil Aggarwal  wrote:
> 
> >
> > Hello Team Pacemaker,
> >
> > I am facing a constant issue with Pacemaker, it does not restart the 
> > Service even when he knows that the Service is down. It generates a message 
> > saying "Ignoring Expired Failure" for the service.
> 
> What is the failure timeout set to?
> 
> > Pacemaker and Corosync version are given below. OS CentOS 6.2
> >
> > corosync-1.4.1-4.el6_2.2.x86_64 pacemaker-1.1.9-2.el6.x86_64
> >
> > Log which pengine provide is:
> >
> >  pengine[45232]:   notice: unpack_rsc_op: Ignoring expired failure 
> > (calculated) Server_last_failure_0 (rc=7, 
> > magic=0:7;14:5699:0:459093cc-f3a1-483b-b853-53a1d9791361)
> >
> > Some more info is:
> >
> > 1.This is a two node cluster. There is time difference of 10 min b/w the 
> > two nodes.
> >
> >
> > --
> > Regards,
> > Sahil
> > Mobile - 09467607999
> > fbAddress-www.facebook.com/SahilAggarwalg
> 
> 
> 
> 
> -- 
> Sahil
> Mobile - 09467607999
> fbAddress-www.facebook.com/SahilAggarwalg



signature.asc
Description: Message signed with OpenPGP using GPGMail
___
Pacemaker mailing list: Pacemaker@oss.clusterlabs.org
http://oss.clusterlabs.org/mailman/listinfo/pacemaker

Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://bugs.clusterlabs.org

Re: [Pacemaker] Linux HA setup for CentOS 6.5

2014-10-15 Thread Digimer 

Let pacemaker start cman/corosync on EL6.

This is the guide that covers it, written by Pacemaker's author:

http://clusterlabs.org/doc/en-US/Pacemaker/1.1-plugin/html-single/Clusters_from_Scratch/index.html

It notes that it's based on Fedora 13, but that maps to EL6 almost 
perfectly.


A very slightly altered approach is here, in my *very* unfinished tutorial:

https://alteeve.ca/w/Anvil!_Tutorial_3_on_EL6#Configuring_the_Anvil.21

The main difference is that Andrew's approach (see section 8.2.2) is to 
disable quorum via editing /etc/sysconfig/cman, where my approach 
handles it in the main /etc/cluster/cluster.conf (cman's main config file).


In any case, from then on, start pacemaker and let it handle everything 
else.


Cheers

digimer

On 15/10/14 04:27 AM, Sihan Goi wrote:

Hi,

So I've decided to make things simpler and go with a wired network
instead of wireless. I connected both boxes to a router, manually edited
the ifcfg-eth0 files to set static IP addresses for both boxes (not
before downloading and building a driver for the nic of 1 of the boxes),
did a "chkconfig NetworkManager off", "service NetworkManager stop", and
"service network restart".

I'm able to ping each other via IP address and hostname. I also already
have corosync, pacemaker, crmsh and cman installed.

I then did the following as per the guide at
http://geekpeek.net/linux-cluster-corosync-pacemaker

service corosync start - success.
service pacemaker start - I get a "Starting cman...corosync cluster
engine is already running [FAILED]"

What's up? :(

On Oct 15, 2014 12:23 PM, "Sihan Goi" mailto:gois...@gmail.com>> wrote:

No typo.

[root@node02 network-scripts]# ls -lah
/etc/sysconfig/network-scripts/ifcfg-*
-rw-r--r--. 1 root root 254 Oct 10  2013
/etc/sysconfig/network-scripts/ifcfg-lo

I installed CentOS 6.5 with the LiveDVD. I found it weird as well
that these files were missing.

On Wed, Oct 15, 2014 at 11:54 AM, Digimer mailto:li...@alteeve.ca>> wrote:

Sure there isn't a typo there?

an-c05n01:~# ls -lah /etc/sysconfig/network-__scripts/ifcfg-*
-rw-r--r--. 1 root root 225 Jan 16  2013
/etc/sysconfig/network-__scripts/ifcfg-bond0
-rw-r--r--. 1 root root 220 Jan 16  2013
/etc/sysconfig/network-__scripts/ifcfg-bond1
-rw-r--r--. 1 root root 198 Jan 16  2013
/etc/sysconfig/network-__scripts/ifcfg-bond2
-rw-r--r--. 1 root root 149 Jan 16  2013
/etc/sysconfig/network-__scripts/ifcfg-eth0
-rw-r--r--. 1 root root 144 Jan 16  2013
/etc/sysconfig/network-__scripts/ifcfg-eth1
-rw-r--r--. 1 root root 152 Mar 14  2013
/etc/sysconfig/network-__scripts/ifcfg-eth2
-rw-r--r--. 1 root root 149 Jan 16  2013
/etc/sysconfig/network-__scripts/ifcfg-eth3
-rw-r--r--. 1 root root 144 Jan 16  2013
/etc/sysconfig/network-__scripts/ifcfg-eth4
-rw-r--r--. 1 root root 152 Mar 14  2013
/etc/sysconfig/network-__scripts/ifcfg-eth5
-rw-r--r--. 1 root root 254 Jul 22 09:56
/etc/sysconfig/network-__scripts/ifcfg-lo
-rw-r--r--. 1 root root 213 Mar 13  2013
/etc/sysconfig/network-__scripts/ifcfg-vbr2

I've never seen an EL6 install without the files there,
'network' or NetworkManager aside.

digimer

On 14/10/14 11:32 PM, Sihan Goi wrote:

There aren't any config files in
/etc/sysconfig/network-__scripts. When I
was using CentOS 7, the config files were there
(ifcfg-something) but in
this CentOS 6.5 installation, they are missing.

If is possible to not use cman, and just use corosync and
pacemaker? If
so, how?

On Wed, Oct 15, 2014 at 11:22 AM, Digimer mailto:li...@alteeve.ca>
>> wrote:

 You can manually configure the wireless LAN without
NetworkManager.
 If you take a look, there should be existing config
files in
 /etc/sysconfig/network-scripts/ for the wireless
connection. I've
 not done it myself since many Fedora's ago, but I
believe you can
 change NMCONTROLLER="no" and then start it up with
 /etc/sysconfig/network start. I could be a bit wrong,
but I am sure
 you can make wireless work without NM.

 Question; Servers with WLAN? I assume these won't be
used for corosync?

 digimer


 On 14/10/14 11:17 PM, Sihan Goi wrote:

 Hi,

 Is there a tutorial showing how to get a basic
Linux HA setup with
 replicated storage (via DRBD) working on CentOS
6.5? I want to have
 mySQL as the HA resource with

Re: [Pacemaker] Linux HA setup for CentOS 6.5

2014-10-15 Thread Sihan Goi 
Hi,

So I've decided to make things simpler and go with a wired network instead
of wireless. I connected both boxes to a router, manually edited the
ifcfg-eth0 files to set static IP addresses for both boxes (not before
downloading and building a driver for the nic of 1 of the boxes), did a
"chkconfig NetworkManager off", "service NetworkManager stop", and "service
network restart".

I'm able to ping each other via IP address and hostname. I also already
have corosync, pacemaker, crmsh and cman installed.

I then did the following as per the guide at
http://geekpeek.net/linux-cluster-corosync-pacemaker

service corosync start - success.
service pacemaker start - I get a "Starting cman...corosync cluster engine
is already running [FAILED]"

What's up? :(
On Oct 15, 2014 12:23 PM, "Sihan Goi"  wrote:

> No typo.
>
> [root@node02 network-scripts]# ls -lah
> /etc/sysconfig/network-scripts/ifcfg-*
> -rw-r--r--. 1 root root 254 Oct 10  2013
> /etc/sysconfig/network-scripts/ifcfg-lo
>
> I installed CentOS 6.5 with the LiveDVD. I found it weird as well that
> these files were missing.
>
> On Wed, Oct 15, 2014 at 11:54 AM, Digimer  wrote:
>
>> Sure there isn't a typo there?
>>
>> an-c05n01:~# ls -lah /etc/sysconfig/network-scripts/ifcfg-*
>> -rw-r--r--. 1 root root 225 Jan 16  2013 /etc/sysconfig/network-
>> scripts/ifcfg-bond0
>> -rw-r--r--. 1 root root 220 Jan 16  2013 /etc/sysconfig/network-
>> scripts/ifcfg-bond1
>> -rw-r--r--. 1 root root 198 Jan 16  2013 /etc/sysconfig/network-
>> scripts/ifcfg-bond2
>> -rw-r--r--. 1 root root 149 Jan 16  2013 /etc/sysconfig/network-
>> scripts/ifcfg-eth0
>> -rw-r--r--. 1 root root 144 Jan 16  2013 /etc/sysconfig/network-
>> scripts/ifcfg-eth1
>> -rw-r--r--. 1 root root 152 Mar 14  2013 /etc/sysconfig/network-
>> scripts/ifcfg-eth2
>> -rw-r--r--. 1 root root 149 Jan 16  2013 /etc/sysconfig/network-
>> scripts/ifcfg-eth3
>> -rw-r--r--. 1 root root 144 Jan 16  2013 /etc/sysconfig/network-
>> scripts/ifcfg-eth4
>> -rw-r--r--. 1 root root 152 Mar 14  2013 /etc/sysconfig/network-
>> scripts/ifcfg-eth5
>> -rw-r--r--. 1 root root 254 Jul 22 09:56 /etc/sysconfig/network-
>> scripts/ifcfg-lo
>> -rw-r--r--. 1 root root 213 Mar 13  2013 /etc/sysconfig/network-
>> scripts/ifcfg-vbr2
>>
>> I've never seen an EL6 install without the files there, 'network' or
>> NetworkManager aside.
>>
>> digimer
>>
>> On 14/10/14 11:32 PM, Sihan Goi wrote:
>>
>>> There aren't any config files in /etc/sysconfig/network-scripts. When I
>>> was using CentOS 7, the config files were there (ifcfg-something) but in
>>> this CentOS 6.5 installation, they are missing.
>>>
>>> If is possible to not use cman, and just use corosync and pacemaker? If
>>> so, how?
>>>
>>> On Wed, Oct 15, 2014 at 11:22 AM, Digimer >> > wrote:
>>>
>>> You can manually configure the wireless LAN without NetworkManager.
>>> If you take a look, there should be existing config files in
>>> /etc/sysconfig/network-__scripts/ for the wireless connection. I've
>>> not done it myself since many Fedora's ago, but I believe you can
>>> change NMCONTROLLER="no" and then start it up with
>>> /etc/sysconfig/network start. I could be a bit wrong, but I am sure
>>> you can make wireless work without NM.
>>>
>>> Question; Servers with WLAN? I assume these won't be used for
>>> corosync?
>>>
>>> digimer
>>>
>>>
>>> On 14/10/14 11:17 PM, Sihan Goi wrote:
>>>
>>> Hi,
>>>
>>> Is there a tutorial showing how to get a basic Linux HA setup
>>> with
>>> replicated storage (via DRBD) working on CentOS 6.5? I want to
>>> have
>>> mySQL as the HA resource with the database replicated across the
>>> nodes.
>>> I've scoured the web for one but it seems that I get stuck in
>>> each one
>>> somewhere.
>>>
>>> To elaborate, I have 2 CentOS 6.5 nodes configured with distinct
>>> hostnames and static IPs. They are connected to a wireless AP,
>>> and can
>>> ping each other.
>>>
>>> I tried following this guide -
>>> http://clusterlabs.org/__quickstart-redhat.html
>>> 
>>> However, cman will not start when NetworkManager is running, and
>>> my
>>> nodes cannot connect to the wireless AP without NetworkManager
>>> running.
>>> Am I missing something or is that the stupidest dependency ever?
>>> How is
>>> a cluster supposed to work when the nodes aren't connected to
>>> one another?
>>>
>>> I also tried following the "clusters from scratch" guide but
>>> that seems
>>> to rely on systemctl calls which aren't available on CentOS 6.5.
>>>
>>> Any help?
>>>
>>> --
>>> - Goi Sihan
>>> gois...@gmail.com 
>>> >
>>>
>>>
>>> __