Replicated failover working on Linux, but not on Windows

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Replicated failover working on Linux, but not on Windows

Senki
Hi,

I am having trouble to get replicated failover to work on 2 separate
Windows Servers with artemis 2.10.0.

But if I use the same broker.xml configuration files on Linux with
artemis 2.9.0, failover works fine.

One machine (192.168.130.235) is the master and the other one
(192.168.130.236) is the slave. They are on the same LAN and there is no
firewall between them. (Windows firewall is also disabled)

This is the configuration that the master uses:
https://pastebin.com/wMR5yqdy

And this is the configuration that the slave uses:
https://pastebin.com/jAjzE4A9

On Windows, I can start the master and the server works fine, but if I
start the slave, it fails to locate the master. The server doesn't
start, it is waiting for a master to be started.

What I've noticed is, that while the group-addressin the
broadcast-groupsand discovery-groupselement is set to 231.7.7.7.

On the Linux master I can see it listening on the given IP address by
running netstat -an | grep 9876:

udp6       0      0 231.7.7.7:9876         
:::*                              
udp6       0      0 231.7.7.7:9876          :::*

But if I run netstat -an | find "9876" on the Windows master i get:

  UDP    0.0.0.0:9876           *:*
  UDP    0.0.0.0:9876           *:*
  UDP    [::]:9876              *:*
  UDP    [::]:9876              *:*

If I run the two servers on the same Windows machine (after replacing
the IP addresses in one of the config files), then the failover works.
That makes me believe that the problem is not being caused by an error
in the config file.

My question is: Do I have to set up discovery on Windows different than
on Linux? Or anything at all?

If anyone has set up failover or a cluster on 2 (or more) separate
windows machines, please let me know how you got it working.

I am not sure, that the problem is being caused by the group-addres,
that is just my guess.

I can attach the log files, if that helps, let me know if you want to
see them.

All suggestions and help are appreciated.

Regards,

Adam

Reply | Threaded
Open this post in threaded view
|

Re: Replicated failover working on Linux, but not on Windows

jbertram
> Do I have to set up discovery on Windows different than on Linux? Or
anything at all?

You shouldn't have to configure anything differently in broker.xml in
Windows vs. Linux to get cluster discovery to work. Since it is working
with 2 local broker and not working with 2 remote brokers my guess is that
there's something going on with UDP multicast between the 2 remote brokers
where it doesn't work. Sometimes network administrators disable UDP
multicast or severely limit the propagation of the packets. I'd do some
more investigation into your environment.


Justin

On Thu, Sep 19, 2019 at 8:18 AM Senki <[hidden email]> wrote:

> Hi,
>
> I am having trouble to get replicated failover to work on 2 separate
> Windows Servers with artemis 2.10.0.
>
> But if I use the same broker.xml configuration files on Linux with
> artemis 2.9.0, failover works fine.
>
> One machine (192.168.130.235) is the master and the other one
> (192.168.130.236) is the slave. They are on the same LAN and there is no
> firewall between them. (Windows firewall is also disabled)
>
> This is the configuration that the master uses:
> https://pastebin.com/wMR5yqdy
>
> And this is the configuration that the slave uses:
> https://pastebin.com/jAjzE4A9
>
> On Windows, I can start the master and the server works fine, but if I
> start the slave, it fails to locate the master. The server doesn't
> start, it is waiting for a master to be started.
>
> What I've noticed is, that while the group-addressin the
> broadcast-groupsand discovery-groupselement is set to 231.7.7.7.
>
> On the Linux master I can see it listening on the given IP address by
> running netstat -an | grep 9876:
>
> udp6       0      0 231.7.7.7:9876
> :::*
> udp6       0      0 231.7.7.7:9876          :::*
>
> But if I run netstat -an | find "9876" on the Windows master i get:
>
>   UDP    0.0.0.0:9876           *:*
>   UDP    0.0.0.0:9876           *:*
>   UDP    [::]:9876              *:*
>   UDP    [::]:9876              *:*
>
> If I run the two servers on the same Windows machine (after replacing
> the IP addresses in one of the config files), then the failover works.
> That makes me believe that the problem is not being caused by an error
> in the config file.
>
> My question is: Do I have to set up discovery on Windows different than
> on Linux? Or anything at all?
>
> If anyone has set up failover or a cluster on 2 (or more) separate
> windows machines, please let me know how you got it working.
>
> I am not sure, that the problem is being caused by the group-addres,
> that is just my guess.
>
> I can attach the log files, if that helps, let me know if you want to
> see them.
>
> All suggestions and help are appreciated.
>
> Regards,
>
> Adam
>
>
Reply | Threaded
Open this post in threaded view
|

Re: Replicated failover working on Linux, but not on Windows

wazburrows
In reply to this post by Senki
Senki, you said that failover works for you on Linux. Does it also fail back
from the backup to the master again on linux?  The problem I'm seeing is
that artemis 2.10 on Linux will failover to a backup but doesn't fail back
correctly when the primary comes up again.

Thanks




--
Sent from: http://activemq.2283324.n4.nabble.com/ActiveMQ-User-f2341805.html