Cluster or other redundancy

Posted on
Wed Oct 01, 2014 7:15 am
lochnesz offline
Posts: 370
Joined: Oct 01, 2014
Location: Stockholm, Sweden

Cluster or other redundancy

I'm currently using HC2, Vera Lite and Zipabox, but I am getting more and more curious about Indigo. One important issue for me is to be able to have redundancy, in case the controller fails. This is key, because so many functions in my home depends on a working controller (Z-wave in my case). I am not impressed with the way my current controllers handle this.

How can you build a clustered or redundant Indigo system?

Many thanks!

Posted on
Wed Oct 01, 2014 8:55 am
DaveL17 offline
User avatar
Posts: 6759
Joined: Aug 20, 2013
Location: Chicago, IL, USA

Re: Cluster or other redundancy

I would like to hear more about this too. Here's what I think.

Because the routing maps for Z-Wave reside in the controller and devices themselves, if the controller goes down, so does the automation. I don't think it's possible to have multiple Z-Wave controllers on one network.

Many folks have Indigo on a battery backup in case of a power failure (some run it on a laptop with its internal battery.). But if you don't have power, that obviously affects wired devices. I have Indigo on a battery backup and I also have a whole-home natural gas generator. So I just need Indigo to stay up until the generator kicks in.

I also have automation built into Indigo -- just in case. By that, I mean periodic checks to ensure that devices are in the proper state at set times of the day. For example, I have a schedule that runs at midnight which "re-runs" schedules that have run in the last couple of hours. Suspenders and a belt. I also have Growl notices that fire routinely and email notices for emergencies--just in case.

In all, I have found Indigo itself to be incredibly stable and, after I got my setup tuned, find very little that needs attention.

That said, I am careful with certain key things. For example, I'm not comfortable putting my sump pump on a smart switch for fear that the switch could fail in the off position.

I think stability and fault tolerance are a huge part of home automation. Give Indigo a try, you will not be disappointed. I came over from Vera and couldn't be happier.

I came here to drink milk and kick ass....and I've just finished my milk.

[My Plugins] - [My Forums]

Posted on
Wed Oct 01, 2014 1:44 pm
lochnesz offline
Posts: 370
Joined: Oct 01, 2014
Location: Stockholm, Sweden

Re: Cluster or other redundancy

You have taken it quite far with a generator for your house, I only have a UPS keeping essentials alive for about one hour. My main concern isn't power, but hardware failure (i.e. Mac breaks down). I like the suspenders and belts approach though.

Regarding two Z-wave controllers on one network, there are clustering solutions, and there are also definitions with primary and secondary controllers. I wonder if Indigo has implemented any of these, and if the solution would be seamless in case one controller fails.

Posted on
Wed Oct 01, 2014 1:50 pm
kw123 offline
User avatar
Posts: 8366
Joined: May 12, 2013
Location: Dallas, TX

Re: Cluster or other redundancy

there are several discussion round this, look for backup & zwave .... in summary no backup solution.

Just from my side: I am using z-wave with the SiliconLAB USB stick and it has been working with NO problems for > 9 months now.

Karl

Posted on
Wed Oct 01, 2014 3:51 pm
DaveL17 offline
User avatar
Posts: 6759
Joined: Aug 20, 2013
Location: Chicago, IL, USA

Re: Cluster or other redundancy

lochnesz wrote:
You have taken it quite far with a generator for your house, I only have a UPS keeping essentials alive for about one hour. My main concern isn't power, but hardware failure (i.e. Mac breaks down). I like the suspenders and belts approach though.

Regarding two Z-wave controllers on one network, there are clustering solutions, and there are also definitions with primary and secondary controllers. I wonder if Indigo has implemented any of these, and if the solution would be seamless in case one controller fails.
Yeah, the generator was a big purchase. We live in the country, and have been known to lose power for days at a time. The worst was when we lost power for three days--the house got down to 40 F and we had about 10 in of water in the basement. The generator came as soon as we could afford it.

It comes on automatically, and is strong enough to run the entire house--including both central AC units. It has come in handy on several occasions since!

I came here to drink milk and kick ass....and I've just finished my milk.

[My Plugins] - [My Forums]

Posted on
Wed Oct 01, 2014 6:01 pm
jay (support) offline
Site Admin
User avatar
Posts: 18225
Joined: Mar 19, 2008
Location: Austin, Texas

Re: Cluster or other redundancy

Redundancy for Z-Wave controllers is not a consistently implemented feature. The protocol does allow you to have secondary controllers - you push your network info from your single primary (single point of failure) to secondary controllers. Here's the rub however - controllers have to implement the correct firmware to be able to receive and send network information. They also need to have the ability to become the primary controller. The Z-Stick doesn't do this so there really is no way to have automatic backups and recovery.

Indigo makes switching out your interface a bit easier - when get a new interface and include a device, you can define and sync an existing device in Indigo and select the new node it from the new interface. It still means touching each device, but it's better than starting everything over. We're hoping that things will get better eventually, but for now that's the state of things.

Jay (Indigo Support)
Twitter | Facebook | LinkedIn

Posted on
Thu Oct 02, 2014 1:01 am
lochnesz offline
Posts: 370
Joined: Oct 01, 2014
Location: Stockholm, Sweden

Re: Cluster or other redundancy

We're hoping that things will get better eventually, but for now that's the state of things.


Are you at Indigo trying to get Aeon to make changes in their implementation? It is key to have a reliable and redundant system, and my main concern is the Mac hardware. Maybe there can be other creative solutions to get some kind of redundancy? Two Mac's sharing the same USB device? I don't know if that would work, but there are USB sharing possibilities for other products. Then you would get one less single point of failure. Any thoughts?

Posted on
Thu Oct 02, 2014 7:07 am
RogueProeliator offline
User avatar
Posts: 2501
Joined: Nov 13, 2012
Location: Baton Rouge, LA

Re: Cluster or other redundancy

It is key to have a reliable and redundant system, and my main concern is the Mac hardware.

If you look over or follow the forums for a while I think that you'll find the Mac hardware is very rarely the point of failure in an automation system built on Indigo. Not saying it doesn't happen, but by far the most reported/discussed failures seem to be the devices themselves followed by the controllers. Put your Mac on a solid UPS, keep your OS updated, follow standard security practices, back up the system and don't worry too much about the computer hardware.

Adam

Posted on
Thu Oct 02, 2014 10:41 am
jay (support) offline
Site Admin
User avatar
Posts: 18225
Joined: Mar 19, 2008
Location: Austin, Texas

Re: Cluster or other redundancy

lochnesz wrote:
my main concern is the Mac hardware.


That part is quite simple - it's the part of your system that is the least likely to fail. Put your Mac on a UPS and it's up all the time even through power outages. Well, I should say, a dedicated Mac will be the least likely part of the system to fail. Keep a Time Machine backup going on a separate disk and you're solid. I really don't think keeping redundant Macs is going to be cost-effective since it's the least likely to fail. Just my opinion of course.

I'm using a first generation MacBook Pro from 2006 as my home server (and have been for 6 years).

Oh, and I'm agreeing with Adam with one caveat: never EVER upgrade to a new OS release until it's been out at least 6 months if you want to maintain a stable, reliable Indigo Server. Unfortunately, recent history has made me even more cautious about new Mac OS releases (and for that matter iOS releases). I'm sure I'll get all sorts of "but it works for me", but having been a Mac user, developer, and computer support manager for 26 years, I'm just saying that when dealing with a server machine, staying on the bleeding edge is not good for uptime.

Jay (Indigo Support)
Twitter | Facebook | LinkedIn

Posted on
Thu Oct 02, 2014 11:16 am
lochnesz offline
Posts: 370
Joined: Oct 01, 2014
Location: Stockholm, Sweden

Re: Cluster or other redundancy

Jay, Adam, many thanks for your support and comments. I agree with you that the Mac is quite good quality, I currently have 7 of them, but I also had my share of broken Mac's. Currently my HC2 is my main controller and my Vera is secondary which is not a pretty solution, but I can keep my house running if the HC2 fails. They are both on UPS power. To give you a little better picture of my situation, the house is my summer home, 3500 km from my permanent home, and I visit it 3-4 times per year. Currently I have 57 devices (end points), and some of the devices serves as an alarm system. Apart from that, the septic tank pump is on a z-wave switch, and terrace and pool lights are controlled from z-wave too. Since part of the property is rented out during part of the year, and the guests have no access to the terrace light and pool light switches, this just have to work. The same goes for air conditioning control. I have double end points for some devices, so if one end point fails, I can still turn on the AC for example.

My point is, that keeping a backup of the Mac wouldn't be enough, since if the system fails, there is no one there to restore the backup.

I understand what you are saying, I would have to live with a non redundant system if I decide to switch to Indigo. I hope that there will be more redundant solutions in the future, though.

Posted on
Thu Oct 02, 2014 12:02 pm
RogueProeliator offline
User avatar
Posts: 2501
Joined: Nov 13, 2012
Location: Baton Rouge, LA

Re: Cluster or other redundancy

Oh, and I'm agreeing with Adam with one caveat: never EVER upgrade to a new OS release until it's been out at least 6 months if you want to maintain a stable, reliable Indigo Server.

By "keep the OS updated" I actually mean the security updates for the OS that you have on there... I actually wouldn't updated new a new OS full version for at least six months on any critical machines. I'm more likely to upgrade to the "old" version just before a new one comes out. I just installed Mavericks on my development laptop and still haveMountain Lion on my Indigo server (with no plans to upgrade at all for now).

You could certainly mirror your server and have a complete, virtually identical backup if the expense is worth it. However, it sounds as if you would be stuck in terms of the controller...

Adam

Page 1 of 1

Who is online

Users browsing this forum: No registered users and 9 guests