Regular 'apparent' signal loss on ETH6?

Ask fellow members about Ceton's infiniTV tuners here.
Forum rules
Ceton no longer participate in this forum. Official support may still be handled via the Ceton Ticket system.
Post Reply
scyto

Posts: 164
Joined: Thu Sep 06, 2012 7:46 pm
Location:

HTPC Specs: Show details

Regular 'apparent' signal loss on ETH6?

#1

Post by scyto » Sun Apr 03, 2016 4:46 pm

Every so often I will come to media center and find all my scheduled recordings failed. Looking in history shows there was no signal on any tuner. Using the cable card tab of the Ceton diags app shows everything is ok except step9 which reports signal issue. This happens once every two to six weeks or so - i.e. random but regular frequency.

I used to think this was the ETH. But I have found that a reboot of the media center fixes this issue every time and no reboot of the infinitv is required.

At one point I thought this might be due to pending windows or other sw updates, but now I am not so sure.

Any ideas? Should I just workaround with a once a week 3am reboot of the media center? (My preference is to fix not workaround)

User avatar
spanner

Posts: 269
Joined: Mon Jun 06, 2011 6:36 pm
Location: Corbett Oregon USA

HTPC Specs: Show details

#2

Post by spanner » Sun Apr 03, 2016 7:35 pm

Not sure this is related, and i have not had trouble with recordings yet, but i often get this week signal message since comcast switched some channels to MPEG 4. It happens when i go from a mpeg 2 channel to a mepeg 4 channel. If seems to correct itself if i go to a working channel then back to the week signal channel then all is fine. I have been wondering if this will be a problem during a recording switch.

mdavej

Posts: 1477
Joined: Mon Aug 20, 2012 6:52 pm
Location:

HTPC Specs: Show details

#3

Post by mdavej » Sun Apr 03, 2016 8:50 pm

What's happening on the OP's issue is the ETH has a memory leak. So the memory fills up about every 6 weeks, the ETH spontaneously reboots, and WMC looses its connection. Rebooting WMC re-establishes the connection. This bug has been around forever and will never be fixed. So there is no fix, but a possible workaround.

Somebody made a program that monitors the ETH and reboots the appropriate WMC services automatically, if you want to give that a try. Worked pretty well last time I tried it. I don't remember what it was called though. Seems like it came from someone over at avsforum.com.

You can use this web page to monitor it. You'll see the memory creeping up steadily.
http://www.avsforum.com/forum/26-home-t ... -page.html

scyto

Posts: 164
Joined: Thu Sep 06, 2012 7:46 pm
Location:

HTPC Specs: Show details

#4

Post by scyto » Sun Apr 03, 2016 9:43 pm

Thanks! I will try finding the app, if you happen to come across it before I do please link.

scyto

Posts: 164
Joined: Thu Sep 06, 2012 7:46 pm
Location:

HTPC Specs: Show details

#5

Post by scyto » Mon Apr 04, 2016 1:13 am

scyto wrote:Thanks! I will try finding the app, if you happen to come across it before I do please link.
huh one can't edit ones posts - how crap, anywho.

Do you mean this 'program' http://www.avsforum.com/forum/26-home-t ... st34583465

If so it seems to ping the ETH with a high frequency to see if it is unavailable and then is restarts the ehrecvr service. I am wondering if there is a more elegant / reliable way as the posters indicate there may be some unreliability, for example monitoring the storage and memory variable or uptime using wget or app. I 'don't' code but could persuade myself to try... for example look at the uptime field every minute, and if it is a lower number than last time checked restart ehrcvr.

Do you think this is something folks would be interested in?

//192.168.1.81/get_var?i=0&s=diag&v=Host_Memory_Report
//192.168.1.81/get_var?i=0&s=diag&v=Host_Storage_Report
//192.168.1.81/get_var?i=0&s=diag&v=Uptime'

--edit-- nope writing this as a script / app / service is beyond me and the time i am willing to invest to learn enough coding skills :-( wonder if there are folks for hire on the interwebs...

User avatar
Scallica

Posts: 2797
Joined: Mon Jun 06, 2011 7:09 pm
Location: USA!

HTPC Specs: Show details

#6

Post by Scallica » Mon Apr 04, 2016 3:04 am

scyto wrote:huh one can't edit ones posts
Editing is only allowed within 1 hour of posting. If edits are needed beyond 1 hour, contact a staff member.
HTPC Enthusiast / Forum Moderator - TGB.tv Code of Conduct

mdavej

Posts: 1477
Joined: Mon Aug 20, 2012 6:52 pm
Location:

HTPC Specs: Show details

#7

Post by mdavej » Mon Apr 04, 2016 1:48 pm

scyto wrote:
scyto wrote:Do you mean this 'program'...
Yep. That's it.

That's a good idea you have there. But what I've found is the spontaneous reboot is kind of random. For a while I ran a simple task in task manager that would reboot the ETH once a week, then restart WMC services. That kind of worked, but not 100%. I think something like the periodic ping is more bullet proof.

scyto

Posts: 164
Joined: Thu Sep 06, 2012 7:46 pm
Location:

HTPC Specs: Show details

#8

Post by scyto » Mon Apr 04, 2016 2:16 pm

mdavej wrote: I think something like the periodic ping is more bullet proof.
My concern with ping is it is unreliable, there are all sorts of reasons a transitory error could result in no ping, also it means flooding the network with ping traffic for a gap in service that will about 3 to 5 seconds. If the ETH reboots itself looking at uptime variable is 100% reliable approach; it can also be used to detect if there is no web page. Maybe i will see if I can persuade someone at work to build the code framing for me....

I submitted a ticket to Ceton - lets see what they say, theoretically i can do knock on their door if they don't reply :-)

scyto

Posts: 164
Joined: Thu Sep 06, 2012 7:46 pm
Location:

HTPC Specs: Show details

#9

Post by scyto » Thu Apr 07, 2016 3:43 am

They replied with a genetic update your drivers and this is a demonstration of how unreliable MCE is. Has any one seen this with home run prime?

RyC

Posts: 724
Joined: Tue Aug 21, 2012 10:21 pm
Location:

HTPC Specs: Show details

#10

Post by RyC » Thu Apr 07, 2016 3:52 am

There is a weird bug on the Prime that required a power cycle after a while, although it may have been fixed, but in general, the Prime does not suffer through things like this and is more reliable. I personally find the ETH6 to be reliable enough, and the 6 tuners with 1 cablecard worth it

Space

Posts: 2838
Joined: Sun Jun 02, 2013 9:44 pm
Location:

HTPC Specs: Show details

#11

Post by Space » Thu Apr 07, 2016 5:01 am

The Prime has something called the "47 hour DRM bug", but I don't think it is actually a bug with the Prime itself (although I am not sure). You can find info on it here:
http://www.thegreenbutton.tv/forums/vie ... f=6&t=5927.

The Prime has no such other bugs that leave it inoperable as far as I know. It just works every time, all the time.

It may be possible to implement a workaround for the DRM bug in the Prime firmware, and I don't know if any such workaround was ever implemented.

You can avoid the bug by either rebooting the Prime, the HTPC, or putting the HTPC to sleep/awake at least once every 47 hours.

The bug manifests itself by being unable to record any DRM channel starting at exactly the 47 hour mark (47 hours from the last reboot of Prime/HTPC or last wake from standby). If you are in the middle of recording a DRM protected program when the 47 hour timer expires, the video will just go black.

Since I have my HTPC sleep whenever not in use, I "never" see this bug. The only time I saw it was when recording/watching the Olympics 24/7 when the HTPC was not put to sleep for long periods of time over the course of two weeks.

There is a script in the linked thread that reboots the Prime on a regular basis to avoid the bug (only needed if your PC does not go to sleep regularly, and assumes that a workaround has not already been implemented in the later Prime firmwares).

scyto

Posts: 164
Joined: Thu Sep 06, 2012 7:46 pm
Location:

HTPC Specs: Show details

#12

Post by scyto » Thu Apr 07, 2016 5:54 am

Thanks, I have been trying to get sleep to work - machine wakes up 8 seconds after sleeping. I have 2 other network cards on order as I think the onboard Intel is borked.... Any other tips?

Space

Posts: 2838
Joined: Sun Jun 02, 2013 9:44 pm
Location:

HTPC Specs: Show details

#13

Post by Space » Thu Apr 07, 2016 6:03 am

Did you try disabling wake on LAN?

Go to device manager, find your network adapter, go to properties, and in power management uncheck "Allow this device to wake the computer"

User avatar
Scallica

Posts: 2797
Joined: Mon Jun 06, 2011 7:09 pm
Location: USA!

HTPC Specs: Show details

#14

Post by Scallica » Thu Apr 07, 2016 11:37 am

Space wrote:The Prime has something called the "47 hour DRM bug", but I don't think it is actually a bug with the Prime itself (although I am not sure).
The Prime has no such other bugs that leave it inoperable as far as I know. It just works every time, all the time.
I haven't heard of this bug. My guess is that the bug was fixed a long time ago. I am using the latest firmware. I record DRM shows all the time and don't use sleep mode and don't regularly reboot my Prime.
HTPC Enthusiast / Forum Moderator - TGB.tv Code of Conduct

Ed 

Posts: 808
Joined: Sat May 24, 2014 3:14 pm
Location:

HTPC Specs: Show details

#15

Post by Ed  » Thu Apr 07, 2016 3:13 pm

Yeah I never had that issue with the Prime either when I used WMC. But my HTPC is always sleeping when not in use.

Space

Posts: 2838
Joined: Sun Jun 02, 2013 9:44 pm
Location:

HTPC Specs: Show details

#16

Post by Space » Thu Apr 07, 2016 3:34 pm

Scallica wrote:
Space wrote:The Prime has something called the "47 hour DRM bug", but I don't think it is actually a bug with the Prime itself (although I am not sure).
The Prime has no such other bugs that leave it inoperable as far as I know. It just works every time, all the time.
I haven't heard of this bug. My guess is that the bug was fixed a long time ago. I am using the latest firmware. I record DRM shows all the time and don't use sleep mode and don't regularly reboot my Prime.
OK, I misremembered a small detail of this bug...

It only affects a DRM show that is being recorded or watched live if it is being recorded/watched during the 47 hour timer expiring. If you record or watch a show after the timer expires, it will be fine (and the timer is reset again).

So the bug may still exist, it is just a very rare occurrence.

You can test that the bug still exists by rebooting your Prime and then setting WMC to record a DRM show that is airing exactly 47 hours from the time you rebooted. Make sure the expire time occurs somewhere in the middle of the recording and that you don't reboot your HTPC/Prime or put the HTPC to sleep.
You will know the bug still exists if your video goes black (and I believe no audio) at the time of the timer expiring. The video will continue to record, but it will just be a black screen.

User avatar
STC

Posts: 6808
Joined: Mon Jun 06, 2011 4:58 pm
Location:

HTPC Specs: Show details

#17

Post by STC » Thu Apr 07, 2016 4:12 pm

For Ceton, a general thing to check is tuner temps.
By the Community, for the Community. 100% Commercial Free.

Want decent guide data back? Check out EPG123

mdavej

Posts: 1477
Joined: Mon Aug 20, 2012 6:52 pm
Location:

HTPC Specs: Show details

#18

Post by mdavej » Thu Apr 07, 2016 5:52 pm

Also realize that only the ETH has the leak. If you have a slot, the 4 tuner card would be a reliable alternative.

If you can live with 3 tuners, the Prime has a much brighter future.

scyto

Posts: 164
Joined: Thu Sep 06, 2012 7:46 pm
Location:

HTPC Specs: Show details

#19

Post by scyto » Sat Apr 09, 2016 3:32 am

scyto wrote:Thanks, I have been trying to get sleep to work - machine wakes up 8 seconds after sleeping. I have 2 other network cards on order as I think the onboard Intel is borked.... Any other tips?
Yes, it's the wake on pattern match that is the issue. I want the machine to sleep and then be woken up by the extender.

Controlling pattern match seems to be impossible, even though there are APIs to alter the pattern match packets in Windows there seem to be no utils to do this.

scyto

Posts: 164
Joined: Thu Sep 06, 2012 7:46 pm
Location:

HTPC Specs: Show details

#20

Post by scyto » Sun May 01, 2016 6:11 pm

mdavej wrote:Also realize that only the ETH has the leak. If you have a slot, the 4 tuner card would be a reliable alternative.

If you can live with 3 tuners, the Prime has a much brighter future.
I am trying to keep this WMC bubbling along until it becomes clear what the replacement will be. I agree the prime looks to be center of momentum with either SDDVR with Emby or the DVR get channels is developing for their Apple TV app. What's annoying is my ETH6 was stable on 8/8.1for years, no issues.

At the moment I have a concern these issues may be more WMC issues. I added a prime to the mix on Friday to test if it was purely a ceton issue. Reset the EPG because of no live TV issue even after a reboot. I am now getting unexpected failures of recording service in general, no errors about signal loss, I suspect ehreciever is on its last legs for some reason. I am going to see if a scheduled reboot every 24 hrs is a viable workaround and fixes in same way as a manual reboot does. If not I may fdisk ad start again - this WMC original started life as pre-release WMC7 and has been through many many build to build upgrades to get to 8.1 (I had access to daily build for a while there). Only thing stopping me do it straight away is a bunch of DRM content I still need to watch.

Then I can attempt to prove is this a ceton issue or WMC issue in my instance.

Post Reply