HA Reliability

[123]

Inspired by another topic, I thought it might be interesting for everyone to "open their kimono" and describe the reliability of their HA applications. I'm curious to learn if the failures are similar and if they have common causes.

I use Premise and it runs on Win XP on a dedicated PC (P3, 256K RAM, 65 MB HD) connected to a UPS. Overall it runs consistently and reliably. It does not need to be rebooted as a preventive measure (i.e. to circumvent performance degradation, memory leaks, etc).

I haven't found Win XP to be a liability. Automatic Updates is activated, it runs with Windows Firewall but without anti-virus software.

I've written several custom drivers and found Premise to be fairly resilient to my programming errors. However, it is possible to upset its internal web-server if the driver mishandles the UI. Basically, the core engine continues to function but the web-server stops rendering the UI. A reboot solves the problem (and then I need to fix my code).

PS
I'm speculating that tinkering with drivers/plugins/etc causes most instability issues (more than the OS).

 

[electron]

I run XP as well, and have run different packages on this same machine, none of them forced me to reboot the machine. I treat my XP machine as an appliance, so I don't use it as a workstation, and has been optimized for this role.

I actually turned off Automatic Updates, as that's just a disaster waiting to happen. I also don't have Windows Firewall enabled, since I don't see a need for it.

 

[IVB]

Ditto. I'll answer in 2 parts: Windows, and HA apps. My longest proof point is with SageTV/XP, but i'm not sure folks would consider SageTV to be HA - I know I don't.

For Windows, simply put my PCs do not have any reliability issues that aren't self-inflicted because I muck with crap. When I had a standalone SageTV server, I believe I achieved something like 19 months of straight uptime without so much as a "stop SageTV service & restart", because I was too distracted with CQC to screw with it. I had automatic updates turned off, so there was no auto-install there.

For CQC, i've found it to be at that same level - it does not go down, does not require a reboot to be stable.

But here's the bad news:
For CCTV, my CCTV server used to constantly hang/refuse to accept network connections. I had to set it up to have a nightly reboot. Then chatting with CollinR one night in the chat room, I realized that the way my CQC triggers were set up, I had it auto-load the CCTV screen on the kitchen touchpanel on doorbell ring, but left it on that screen. It would spend 8-20 hours pounding away on the CCTV server doing live streaming of the CCTV feed, indeed sending the quad-cam view. I have since changed that trigger to auto-load the CCTV screen, and 2 minutes later to auto-load a non-CCTV screen. Since then the reliability of the CCTV server

My current issue is with my IRA-3 IR receivers & the IP serial server i'm using for them. I have 2, they used to run rock-solid. I had CAT5 wiring back to the wiring closet, life was great. Then I pulled out my PCI-based MOXA serial card, and went with IP serial servers, as I was swapping mobo's on my server and had fewer slots. Now the same IR receiver, same cable, same CQC controlling it, loses connection perhaps every 24 hours. I'm currently using the CCTV server to control it so I can do an automated stop/start.

I just installed a PCIe based serial card in the CCTV server so I can go back to using a direct serial connection, and my suspicion is that it'll return to rock-solid. I have precedence with devices behaving badly with the IP serial servers (ie, i cannot configure the NuVo over the IP serial server or else it'll corrupt it), so there's a plausible basis for this theory.

 

[AutomatedOutlet]

Never have to reboot - I use a hardware controller.

 

[Skibum]

I use XP on the machine running Homeseer.
I virtually never need to reboot unless there is a windows update pushed.
At that time I let Windows handle everything with the updates, and then after reboot Homeseer and VWS start automatically.

 

[Squintz]

The only time I have to reboot is when I am upgrading to a newer version of CQC and when I write my own driver and want it to show up in the list of CQC drivers for the first time. Other than that it just works. I am running CQC on several machines including 2 - XP Pro, Vista, and XP MCE.

 

[nightwalker]

I used to have to reboot fairly frequently. Iit turned out the problem was self inflicted. Ever since I moved HomeSeer to a dedicated machine it's been running non stop, for months and months, unless I take it down for maintenance of one kind or another. I'm not sure if the key was putting it on a dedicated machine or that I was no longer playing with it, probably the later. :wacko:

 

[IVB]

The only time I have to reboot is when I am upgrading to a newer version of CQC and when I write my own driver and want it to show up in the list of CQC drivers for the first time. Other than that it just works. I am running CQC on several machines including 2 - XP Pro, Vista, and XP MCE.

Why would you reboot? When you upgrade or add a new driver, you need to stop the CQC service anyhow, which is all it needs. I've gone through many CQC upgrades without rebooting that SageTV/CQC box, mainly cuz I do it whilst SageTV is recording or playing something for the wife and she'd kick my ass if I restarted it mid-stream.

 

[123]

I guess I should've used the term "restart" rather than "reboot". Restarting the HA service in order to correct a problem qualifies as a reliability problem.

Restarting the service in order to install a driver doesn't qualify as a reliability concern. However, it'd be different if one was attempting to maximize uptime. Fortunately, most homes can easily accomodate a little downtime!

FWIW, I just restarted the Premise app in an attempt to resolve a connection problem with 3COM Audreys. Premise wasn't the problem ... the Audreys were just cranky and needed to be restarted.

 

[IVB]

I guess I should've used the term "restart" rather than "reboot". Restarting the HA service in order to correct a problem qualifies as a reliability problem.

Doesn't affect my answer, with the exception of the IRA-3 issue i'm currently having which I believe is due to using an IP serial server. Even a hardware based controller would be susceptible to that, as that's "outside the controller box", so it has no bearing on the original question.

 

[ginigma]

I'm running PowerHome on Win XP on rather generic hardware (Asus motherboard). It is very stable because it's only a HA server. I don't have much else running on it. The only time I need to reboot is after patch Tuesday, if one of those patches requires it. I don't have automatic updates turned on. I will update manually.

That said, I usually reboot when I notice the uptime has been 2-3 months, "just because". However, my Linux box now has an uptime of 284 days.