david at lang
Apr 19, 2012, 4:44 PM
Post #10 of 10
(224 views)
Permalink
|
No, heartbeat can work with many nodes.
The original heartbeat v1 only worked with 2 nodes, and the config syntax
(haresources) is limited to that. As of heartbeat v2 it also supports a
xml config that allows you to have many nodes.
If you want to load balance across multiple nodes, you would have
heartbeat use CLUSTERIP to distribute the traffic across the nodes, and if
a node fails, heartbeat will redirect that traffic to another node.
By default, it avoids disrupting existing connections, so if you have
three nodes they would be 1/3, 2/3, 3/3 and then if node 3 fails you would
have 1/2 3/3, 2/3 (one box getting 2/3 of the traffic), I remember seeing
discussion of an option to redistribute the traffic so it would become
1/2, 2/2 instead, but I don't remember if that got accepted or not.
David Lang
On Thu, 19 Apr 2012, Murphy, Brandon wrote:
> Out of curiosity Is heartbeat limited to only 2 nodes? Can more than
> one node be active? There are some of the features that I enjoy about
> having an external load balancer.
>
>
> Brandon Murphy | IT Network Security Assoc-Senior | Information Services - Information Security | The Principal Financial Group? | ph 515.247.5161
>
>
> -----Original Message-----
> From: rsyslog-bounces [at] lists [mailto:rsyslog-bounces [at] lists] On Behalf Of david [at] lang
> Sent: Thursday, April 19, 2012 2:18 PM
> To: rsyslog-users
> Subject: Re: [rsyslog] rsyslog | VIP
>
> You can use heartbeat (http://linux-ha.org) running on your rsyslog servers to manage a VIP on the servers themselves (no need for an external load balancer device)
>
> The hard part of the question comes when you need to determine under what conditions you want to declare a box dead to force a failover.
>
> From a practical point of view, rsyslog is pretty darn reliable, so if you just run rsyslog on both systems and only configure heartbeat to failover if a box is completely dead you will be in pretty good shape.
>
> Hoever, this would not work in the case where rsyslog stops on a box that is otherwise healthy. How much checking you want to do to decide if rsyslog is healthy can be complex, various tests can include:
>
> 1. is rsyslog running (simple ps |grep check)
>
> 2. is rsyslog listening to a port (lsof check, netstat check, or probing a
> port)
>
> 3. is rsyslog processing the message (sec watching for a message from rsyslog on a file or named pipe)
>
> 4. internal status checking in rsyslog
>
> There are also other subtle prroblems that a box can have that will make a box still think it's healthy, but not be working properly for your logging
>
> For example:
>
> Routing problems could cause a box to still be up, but be unable to talk to boxes on other networks
>
> Disk I/O problems could cause the box to be unable to write logs, even though it can still receive them (or if the box is a relay, a downstream box could be not responding so it can't forward the log messages)
>
> If a box runs out of memory, various things on the box will get killed, which may leave a box not working the way it should
>
> I have had about a hundred pairs of boxes running heartbeat for the last decade (more at some points than at others, but averaging around a hundred
> pairs) and I've only had about a dozen or so problems stranger than what the simplist heartbeat configuration detects (full box down issues) during that time, which is why I say that as a practical matter, you are probably just fine with the simplest configuration, without trying to monitor rsyslog explicitly.
>
> David Lang
>
>
>
> On Thu, 19 Apr 2012, Martinez, Carlos R wrote:
>
>> Date: Thu, 19 Apr 2012 13:47:39 +0000
>> From: "Martinez, Carlos R" <Carlos.R.Martinez [at] schwab>
>> Reply-To: rsyslog-users <rsyslog [at] lists>
>> To: "rsyslog [at] lists" <rsyslog [at] lists>
>> Subject: [rsyslog] rsyslog | VIP
>>
>> I'm looking for the best way to put my rsyslog servers behind a VIP for TCP & UDP traffic.
>>
>> It appears that the best solution thus far at least for TCP is fault tolerant setup with port probing. I have tried a VIP load balance/round-robin configuration but it takes a while for the syslog/rsyslog client to start logging to the surviving rsyslog server when I bring one down. As for UDP I'm not having good results at all and I'm starting to wonder if the best solution would be to have a static VIP IP where I use CLUSTERIP configuration on my systems where the failover occurs over the virtual network interface.
>>
>> Any suggestions is appreciated.
>>
>>
>> Here is my configuration:
>>
>>
>> syslog/rsyslog client ->->->-> VIP ->->->-> rsyslog_server1,
>> rsyslog_server2
>>
>> VIP options: 1) fault tolerant w/port probing 2) load balance/round robin 3) static IP (multicast) utilizing CLUSTERIP option.
>>
>> I prefer the load balance configuration for data center HA.
>> _______________________________________________
>> rsyslog mailing list
>> http://lists.adiscon.net/mailman/listinfo/rsyslog
>> http://www.rsyslog.com/professional-services/
>> What's up with rsyslog? Follow https://twitter.com/rgerhards
>>
> _______________________________________________
> rsyslog mailing list
> http://lists.adiscon.net/mailman/listinfo/rsyslog
> http://www.rsyslog.com/professional-services/
> What's up with rsyslog? Follow https://twitter.com/rgerhards
>
>
> -----Message Disclaimer-----
>
> This e-mail message is intended only for the use of the individual or
> entity to which it is addressed, and may contain information that is
> privileged, confidential and exempt from disclosure under applicable law.
> If you are not the intended recipient, any dissemination, distribution or
> copying of this communication is strictly prohibited. If you have
> received this communication in error, please notify us immediately by
> reply email to Connect [at] principal and delete or destroy all copies of
> the original message and attachments thereto. Email sent to or from the
> Principal Financial Group or any of its member companies may be retained
> as required by law or regulation.
>
> Nothing in this message is intended to constitute an Electronic signature
> for purposes of the Uniform Electronic Transactions Act (UETA) or the
> Electronic Signatures in Global and National Commerce Act ("E-Sign")
> unless a specific statement to the contrary is included in this message.
>
> While this communication may be used to promote or market a transaction
> or an idea that is discussed in the publication, it is intended to provide
> general information about the subject matter covered and is provided with
> the understanding that The Principal is not rendering legal, accounting,
> or tax advice. It is not a marketed opinion and may not be used to avoid
> penalties under the Internal Revenue Code. You should consult with
> appropriate counsel or other advisors on all matters pertaining to legal,
> tax, or accounting obligations and requirements.
>
> _______________________________________________
> rsyslog mailing list
> http://lists.adiscon.net/mailman/listinfo/rsyslog
> http://www.rsyslog.com/professional-services/
> What's up with rsyslog? Follow https://twitter.com/rgerhards
>
|
