Controller loosing time

rsw686

Active Member
Since migrating my Zimbra server from CentOS 5 with Zimbra 7.2 to CentOS 6 with Zimbra 8 the OmniPro immediately started loosing exactly 24 seconds every hour. It doesn't make sense that the controller is that sensitive. I'll have to do some traffic traces to see if the new server is sending out some sort of broadcast traffic.
 
What are you running (HAI stuff) on the Linux boxes?
 
Here moving my HAI network port over to the little microrouter (TP-Link) running a modded OpenWRT and manually checking on the time using the old HAI Logger (running in linux - testing still), visuals via PC Access or Homeseer time set I am not seeing this with Firmware 3.14.  The TP-Link Microrouter has an only 8Mb flash and 32Mb RAM running on a 400Mhz CPU.  You can up the 8Mb flash to 32Mb flash if you want.  (also has built in wireless and USB ports).  I am waiting still to update it with an RTC clock having ordered a daughter board for it. 
 
That and I do utilize the GPS configured on the PFSense for NTP time (have shut off any NTP access to the internet).
 
I have a few linux VM (webserver, mail server, etc). The time issue wasn't a problem until I upgraded Zimbra. However it might have just pushed it over the edge. I've done some wireshark captures of broadcast and multicast traffic and my Samsung SmartTV is sending out ARP requests every few seconds. Some searching turns out this is a common occurrence with these TVs. I know the OmniPro doesn't handle large amounts of broadcast. I think I'll just put it on its own VLAN and see if the issue stops.
 
I think I'll just put it on its own VLAN and see if the issue stops.
 
It will not.  I have already tried this.  Thinking I documented here somewhere. 
 
Thinking someone else that purchased a new OPII having network issues sent their board back and it was fixed.  I am curious what they did to fix it as I was told that the issue couldn't be fixed by HAI CS.
 
Best though to try it yourself anyways and tell us your results.
 
pete_c said:
It will not.  I have already tried this.  Thinking I documented here somewhere. 
 
Thinking someone else that purchased a new OPII having network issues sent their board back and it was fixed.  I am curious what they did to fix it as I was told that the issue couldn't be fixed by HAI CS.
 
Best though to try it yourself anyways and tell us your results.
 
The good thing is it has done it consistently over last 14 hours so it will be easy to tell when it is resolved. It used to be off by 10 seconds every 3-4 days. My Vyatta router has an interface with an IP address untagged for my LAN traffic. I added a virtual interface to that port tagged with the VLAN and an IP address on a different subnet. The HAI controller is the only device in this subnet and the switch port the controller connects to is set untagged for this VLAN. This essentially isolates it from broadcast traffic from all other devices. All traffic from my network and Internet must flow through the Vyatta router to reach the HAI controller. If this doesn't resolve it I'm at a loss.
 
Curious about your set up.  Can you draw a stick figure including
 
1 - HAI ethernet port - IP stuff
2 - managed switch connection and configuration of specific ethernet port utilized
3 - specific router / firewall ethernet port configuration (one separate vlan network and another network connection which is untagged).
 
How many ethernet ports do you have configured on your Vyatta router? 
 
Here to test this stuff and other stuff built a new PFSense firewall with multiple Gb nics.
 
 
Well it's not a broadcast traffic issue. Its still loosing exactly 24 seconds an hour. I haven't changed the automation rules in weeks. It happened around midnight last night. Probably will close off internet access to it next to see if that makes a change. Other than the time issue everything is working correctly. I'll put a drawing together tonight or tomorrow.
 
Thank-you rsw686.
 
Curious how you are looking at the time and how you are setting the time and if you have adjusted the time offset in PCA?
 
It's still loosing time with the port closed to the Internet. I'm using my C# logger and the HAI C# SDK. I'm currently flashing the firmware from 3.13 to 3.14a and will resend the PC Access configuration to the controller. It looks like it was a coincidence that it happened the same time I upgraded my linux box.
 
11/21/2014 04:57:00 WARN: TimeSyncTimer: Controller time 11/21/2014 04:56:50 out of sync by 10.2056536 seconds
11/24/2014 20:24:00 WARN: TimeSyncTimer: Controller time 11/24/2014 20:23:50 out of sync by 10.1015933 seconds
11/28/2014 12:51:00 WARN: TimeSyncTimer: Controller time 11/28/2014 12:50:50 out of sync by 10.1139834 seconds
12/02/2014 06:19:00 WARN: TimeSyncTimer: Controller time 12/02/2014 06:18:50 out of sync by 10.081509 seconds
12/05/2014 23:47:00 WARN: TimeSyncTimer: Controller time 12/05/2014 23:46:50 out of sync by 10.1672358 seconds

12/06/2014 00:48:00 WARN: TimeSyncTimer: Controller time 12/06/2014 00:47:36 out of sync by 24.125924 seconds
12/06/2014 01:49:00 WARN: TimeSyncTimer: Controller time 12/06/2014 01:48:36 out of sync by 24.1146523 seconds
12/06/2014 02:50:00 WARN: TimeSyncTimer: Controller time 12/06/2014 02:49:36 out of sync by 24.1816733 seconds
12/06/2014 03:51:00 WARN: TimeSyncTimer: Controller time 12/06/2014 03:50:36 out of sync by 24.0899446 seconds
12/06/2014 04:52:00 WARN: TimeSyncTimer: Controller time 12/06/2014 04:51:36 out of sync by 24.168958 seconds
12/06/2014 05:53:00 WARN: TimeSyncTimer: Controller time 12/06/2014 05:52:36 out of sync by 24.1235902 seconds
12/06/2014 06:54:00 WARN: TimeSyncTimer: Controller time 12/06/2014 06:53:36 out of sync by 24.1554476 seconds
12/06/2014 07:55:00 WARN: TimeSyncTimer: Controller time 12/06/2014 07:54:36 out of sync by 24.1260376 seconds
12/06/2014 08:56:00 WARN: TimeSyncTimer: Controller time 12/06/2014 08:55:36 out of sync by 24.1606325 seconds
12/06/2014 09:57:00 WARN: TimeSyncTimer: Controller time 12/06/2014 09:56:36 out of sync by 24.1807877 seconds
 
Thank you from back in 2012 you provided to me your logger because of the time issues I had with firmware 3.10A.
 
Time Issues
 
I used your logger program to check the time loss.  It was an odd issue where the panel would work fine for months then the issue would occur randomly.
 
Back then it was a gradual then fast time sync loss.  From that I would get serial comm issues which I would see in the Omnistat.  From the serial comm issues I would then see issues with the network port.
 
I attributed the time sync loss issues to the network port.  Tested a disconnect of the Ethernet port and the issue would go away.  This wasn't practical though. 
 
I did get the issue to go away.  That said when I upgraded the firmware to 3.14A it started some new issues.  I basically just cold booted the OPII panel and reloaded my stuff via a serial PCA connection which worked for me. 
 
So far it looks like flashing the firmware and uploading the config from PC Access has fixed the issue. I'll have to wait until tomorrow to call it a success. Either way this shouldn't of happened. I'm going to leave it on its separate network segment for now.
 
Here's a quick diagram of the network layout and config settings. This is your basic router on a stick setup. Traffic between VLAN 1 and 3 always has to pass through the router. I am using VyOS, which is the open source successor to Vyatta after Brocade bought them.
 
Ideally you wouldn't use VLAN 1 and would have either untagged or tagged traffic per port. I wanted to test this scenario quickly without downtime to any other devices. I work in the IT field so when I get home I just want a simple setup that works, which is way it was a flat network.
 
omninetwork.jpg
 
I had a weird bouncing panel issue with firmware 3.13.  Firmware 3.12 and prior was doing the time loss, network and serial issues. 
 
Around that time did separate the HAI network from the main house network using the managed switches and that didn't fix my issues.  What fixed it for me was to installing a microrouter between the OPII NIC and the rest of the network.  That said I have modded the router a bit and looking to provide some more functionality on it.  (email, time sync etc).  I also tested same / similiar configuration using an autonomous NIC on the PFSense firewall which worked.
 
I am currently at firmware 3.14 and have not had any issues.  I have left the microrouter in place and waiting on installing an RTC clock to the microrouter.
 
I a using a custom daughter board from Australia made for the microrouter.  Still waiting on the board.  The board will connect to an little RTC.  Access to the OPII panel works via a serial connection and the IP connecting to the board.  That and adding a little 3G/4G dongle would allow some out of band management.
 
 
I'm using my C# logger and the HAI C# SDK
 
Curiosity question:
 
Are you running your C# logger and HAI C# SDK in Linux or Wintel these days?
 
I have tested your HAI Logger running it in Ubuntu Linux talking to one sql DB on another Ubuntu machine.  Works fine in my round about way of getting it to run.
 
Back
Top