Discussion:
Graylog2 HA
(too old to reply)
c***@public.gmane.org
2014-06-18 12:22:55 UTC
Permalink
Hello,

I thought, I could setup an HA-Graylog2-Cluster. But there has to be only a
single master node, otherwise, I get this message:

Multiple graylog2-server masters in the cluster a few seconds ago
There were multiple graylog2-server instances configured as master in your
Graylog2 cluster. The cluster handles this automatically by launching new
nodes as slaves if there already is a master but you should still fix this.
Check the graylog2.conf of every node and make sure that only one instance
has is_master set to true. Close this notification if you think you
resolved the problem. It will pop back up if you start a second master node
again.

If I ignore this message, things seem to work - but if the master node
dies, you can't login to web-interface anymore :-(

Is there a way to have "real" High-Availability?

regards

Cornelius
--
You received this message because you are subscribed to the Google Groups "graylog2" group.
To unsubscribe from this group and stop receiving emails from it, send an email to graylog2+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/***@public.gmane.org
For more options, visit https://groups.google.com/d/optout.
Martin René Mortensen
2014-06-19 09:06:09 UTC
Permalink
You dont need more than 1 master. The only thing the master does is run
maintenance jobs - everything else runs just as fine on a graylog2 slave.

The web-interface must be configured with only 1 graylog2-server -
otherwise it should work - havent tested this myself.

/Martin
Post by c***@public.gmane.org
Hello,
I thought, I could setup an HA-Graylog2-Cluster. But there has to be only
Multiple graylog2-server masters in the cluster a few seconds ago
There were multiple graylog2-server instances configured as master in your
Graylog2 cluster. The cluster handles this automatically by launching new
nodes as slaves if there already is a master but you should still fix this.
Check the graylog2.conf of every node and make sure that only one instance
has is_master set to true. Close this notification if you think you
resolved the problem. It will pop back up if you start a second master node
again.
If I ignore this message, things seem to work - but if the master node
dies, you can't login to web-interface anymore :-(
Is there a way to have "real" High-Availability?
regards
Cornelius
--
You received this message because you are subscribed to the Google Groups "graylog2" group.
To unsubscribe from this group and stop receiving emails from it, send an email to graylog2+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/***@public.gmane.org
For more options, visit https://groups.google.com/d/optout.
c***@public.gmane.org
2014-06-20 20:09:27 UTC
Permalink
Hi Martin,

sorry, thats not true. If master-node isn't available, at least I have
problems to watch/manage inputs, sometimes I had problems to do a fresh
login, currently I can't reproduce the latter anymore...
This is what I get, if master-node is unavailable:


regards,

Cornelius
Post by c***@public.gmane.org
Hello,
I thought, I could setup an HA-Graylog2-Cluster. But there has to be only
Multiple graylog2-server masters in the cluster a few seconds ago
There were multiple graylog2-server instances configured as master in your
Graylog2 cluster. The cluster handles this automatically by launching new
nodes as slaves if there already is a master but you should still fix this.
Check the graylog2.conf of every node and make sure that only one instance
has is_master set to true. Close this notification if you think you
resolved the problem. It will pop back up if you start a second master node
again.
If I ignore this message, things seem to work - but if the master node
dies, you can't login to web-interface anymore :-(
Is there a way to have "real" High-Availability?
regards
Cornelius
--
You received this message because you are subscribed to the Google Groups "graylog2" group.
To unsubscribe from this group and stop receiving emails from it, send an email to graylog2+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/***@public.gmane.org
For more options, visit https://groups.google.com/d/optout.
Fábio Tramasoli
2015-01-08 16:11:29 UTC
Permalink
Hello, Corneluis.

AFAIK, Martin is right (ref.:
https://www.graylog2.org/resources/documentation/setup/server). My setup
has only one master and it works. Maybe you're only pointing at one
graylog2-server on your graylog2-web install.
I think if the master goes offline another node will be elected as master,
but I'm not sure of this.
Post by c***@public.gmane.org
Hi Martin,
sorry, thats not true. If master-node isn't available, at least I have
problems to watch/manage inputs, sometimes I had problems to do a fresh
login, currently I can't reproduce the latter anymore...
regards,
Cornelius
Post by c***@public.gmane.org
Hello,
I thought, I could setup an HA-Graylog2-Cluster. But there has to be only
Multiple graylog2-server masters in the cluster a few seconds ago
There were multiple graylog2-server instances configured as master in
your Graylog2 cluster. The cluster handles this automatically by launching
new nodes as slaves if there already is a master but you should still fix
this. Check the graylog2.conf of every node and make sure that only one
instance has is_master set to true. Close this notification if you think
you resolved the problem. It will pop back up if you start a second master
node again.
If I ignore this message, things seem to work - but if the master node
dies, you can't login to web-interface anymore :-(
Is there a way to have "real" High-Availability?
regards
Cornelius
--
You received this message because you are subscribed to the Google Groups "graylog2" group.
To unsubscribe from this group and stop receiving emails from it, send an email to graylog2+***@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Pete GS
2015-01-08 21:49:14 UTC
Permalink
We have the same issue here, we actually have 4 Graylog2 servers, two are
dedicated search nodes for the web interface (one of which is the master)
and the other two are dedicated for inputs.

If I ever have the master node offline, the web interface reports it cannot
contact a Graylog2 master node and users cannot log in nor use the
interface, however inputs of course keep accepting data and streams are
still processed.

I have attached a screen dump of the error.

What would be ideal is if there was a similar setting like Elasticsearch
where you can elect nodes that *can* be masters and an election is
performed automatically by the Graylog2 service.
Post by c***@public.gmane.org
Hello,
I thought, I could setup an HA-Graylog2-Cluster. But there has to be only
Multiple graylog2-server masters in the cluster a few seconds ago
There were multiple graylog2-server instances configured as master in your
Graylog2 cluster. The cluster handles this automatically by launching new
nodes as slaves if there already is a master but you should still fix this.
Check the graylog2.conf of every node and make sure that only one instance
has is_master set to true. Close this notification if you think you
resolved the problem. It will pop back up if you start a second master node
again.
If I ignore this message, things seem to work - but if the master node
dies, you can't login to web-interface anymore :-(
Is there a way to have "real" High-Availability?
regards
Cornelius
--
You received this message because you are subscribed to the Google Groups "graylog2" group.
To unsubscribe from this group and stop receiving emails from it, send an email to graylog2+***@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Mike Daoust
2015-02-24 22:52:38 UTC
Permalink
I know this is an old post but I am interested in this behavior as well.
Its problematic to explain that your HA cluster is mostly still working.
Post by c***@public.gmane.org
Hello,
I thought, I could setup an HA-Graylog2-Cluster. But there has to be only
Multiple graylog2-server masters in the cluster a few seconds ago
There were multiple graylog2-server instances configured as master in your
Graylog2 cluster. The cluster handles this automatically by launching new
nodes as slaves if there already is a master but you should still fix this.
Check the graylog2.conf of every node and make sure that only one instance
has is_master set to true. Close this notification if you think you
resolved the problem. It will pop back up if you start a second master node
again.
If I ignore this message, things seem to work - but if the master node
dies, you can't login to web-interface anymore :-(
Is there a way to have "real" High-Availability?
regards
Cornelius
--
You received this message because you are subscribed to the Google Groups "graylog2" group.
To unsubscribe from this group and stop receiving emails from it, send an email to graylog2+***@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Continue reading on narkive:
Loading...