HDHomeRun Prime - constant disconnects

[KryptoNyte]

I have one location with (2) HDHR Prime's with cablecard that are having a nightmare of a time with Prime network disconnects. Anyone else having this issue?

Here's my post;
https://forum.silicondust.com/forum/...?f=113&t=63787

Here's another post by folks that appear to know a fair amount more than I do that are experiencing the same problem and aren't getting much help;
https://forum.silicondust.com/forum/...?f=113&t=63510
(I see NYPlayer posting over there)

This is really frustrating as I've been chasing it for YEARS!

[KryptoNyte]

Do you think if I were to purchase (2) Intel NIC's to add to this machine, and plug the Prime's right into the SageTV Server that it might resolve this problem?

I'm not going to lie, I'm getting tired of just purchasing new equipment to rule out problems with SiliconDust's hardware.

[EnterNoEscape]

Funny you're talking about this today. I had a UPS failure yesterday and my DHCP server came up later than my HDHomeRun devices. Guess what, I just noticed tonight that none of them are reachable. I had to powercycle all of them to get them to pull an IP address. What I found most interesting is that they stayed in this configuration for over 24 hours.

I really wonder if the problem people are seeing is actually the HDHomeRun tries to get/renew its IP address, the DHCP server doesn't respond or the HDHomeRun doesn't get the response within a certain amount of time and then the HDHomeRun statically assigns itself an IP address and turns itself into a DHCP server assuming that it's directly connected to a computer never to try and get an IP address again.

[nyplayer]

Originally Posted by EnterNoEscape

Funny you're talking about this today. I had a UPS failure yesterday and my DHCP server came up later than my HDHomeRun devices. Guess what, I just noticed tonight that none of them are reachable. I had to powercycle all of them to get them to pull an IP address. What I found most interesting is that they stayed in this configuration for over 24 hours.

I really wonder if the problem people are seeing is actually the HDHomeRun tries to get/renew its IP address, the DHCP server doesn't respond or the HDHomeRun doesn't get the response within a certain amount of time and then the HDHomeRun statically assigns itself an IP address and turns itself into a DHCP server assuming that it's directly connected to a computer never to try and get an IP address again.

One of the reasons I got a power strip that Pings my network devices and recycles them after 5 ping fails. I also have reserved IP addresses on all my HdHomerun devices. Also pinging them keeps them active on the Network.

[KryptoNyte]

Originally Posted by nyplayer

One of the reasons I got a power strip that Pings my network devices and recycles them after a set # of ping fails. I also have reserved IP addresses on all my HdHomerun devices.

Awww man ... this is not good news. We're running out of cablecard tuners that actually work. Or rather, we have never had one that could work correctly.

[nyplayer]

Originally Posted by KryptoNyte

Awww man ... this is not good news. We're running out of cablecard tuners that actually work. Or rather, we have never had one that could work correctly.

This is the Strip I got it is well worth the price.

http://www.digital-loggers.com/lpc.html

https://www.amazon.com/Web-Power-Swi...Switch+7&psc=1

[EnterNoEscape]

This situation is exactly why I like to be able to set static IP addresses on my "critical" devices.

[nyplayer]

Originally Posted by EnterNoEscape

This situation is exactly why I like to be able to set static IP addresses on my "critical" devices.

I do both on my Critical Devices. Reserve and Static IP. Attached is my Powerstrip configs.

[JustFred]

Looking at the SiliconDust thread, there's some indication that plugging the Prime into a switch increases the frequency of failure. Is anyone experiencing the failure when their Prime is NOT plugged into a gigabit switch?

Many of the newer gigabit ethernet switches implement "GreenNet" (also known by other names) which reduces the amplitude of signal driving the ethernet cable when there's a short distance between endpoints (i.e., significantly less than the 100 meter max). Unfortunately, this doesn't always work correctly and the result is that one of the two endpoints will decide the link has gone down, so traffic stops. I've had many such reports of this occurring under various combinations of cable length, switch type and end-device.

It would be useful to eliminate the possibility of a link-level GreenNet problem. As a test, try bypassing the switch and plug the Prime directly into the router.

If that's not possible, try inserting a 10/100 switch between the Prime and gigabit switch. Place the 10/100 switch physically near the gig switch, with a relatively short patch cord (< 10 ft should be ok). Most older 10/100 switches don't implement GreenNet, so this trick should take GreenNet out of the equation.

Edit: I have a Prime and 4 other non-gigabit HDHR tuners connected in my network. The Prime is connected via 10 ft patch cord to a TrendNet TEG-S24Dg gigabit switch. Been running this setup for years without any issues. On the other hand, there have been a few customer reviews on Amazon that **hate** this switch, due to the GreenNet problems that were encountered in their specific environment. YMMV.

[nyplayer]

I have mine connected to a NETGEAR ProSAFE 24 port 10/100/1000Mbps Gigibit Ethernet ... and really do not have a problem with them dropping of the network I have 2 primes and 4 connects attached to that switch. Checking for GreenNet is a great Idea.

[EnterNoEscape]

Originally Posted by nyplayer

I do both on my Critical Devices. Reserve and Static IP. Attached is my Powerstrip configs.

I understand what you did there, and I can do reserved IP addresses with my DHCP server to powercycle them when they aren't responding (which doesn't actually happen to me short of the recent power outage), but this whole problem wouldn't even exist if you could actually set a real static IP address.

[JustFred]

Originally Posted by EnterNoEscape

I had a UPS failure yesterday and my DHCP server came up later than my HDHomeRun devices. I had to powercycle all of them to get them to pull an IP address.

This can easily happen if the DHCP server isn't up and running well before any of the clients start making DHCP requests. I avoid the problem by using my programmable power strip to sequence the devices with a 60-second delay between steps: 1: modem, 2: router (dhcp server), 3: server.

Works like a dream.

[nyplayer]

Originally Posted by JustFred

This can easily happen if the DHCP server isn't up and running well before any of the clients start making DHCP requests. I avoid the problem by using my programmable power strip to sequence the devices with a 60-second delay between steps: 1: modem, 2: router (dhcp server), 3: server.

Works like a dream.

That is exactly how I sequence after a power failure.

[EnterNoEscape]

Originally Posted by JustFred

This can easily happen if the DHCP server isn't up and running well before any of the clients start making DHCP requests. I avoid the problem by using my programmable power strip to sequence the devices with a 60-second delay between steps: 1: modem, 2: router (dhcp server), 3: server.

Works like a dream.

No arguments there. I have a similar setup. The issue was that the primary DHCP and DNS server did not turn back on because the power fluctuated too quickly and it ended up staying off.

[KryptoNyte]

Is there a bandwidth concern with having (2) Primes connected to a single 100 megabit switch (then connected back to the gigabit switch)? In the original thread that I linked to at SiliconDust, the response there was to hook the 10/100 switch directly to an extra ethernet port on the server.

[JustFred]

The results of running out of bandwidth will be obvious: dropped packets. But not an unresponsive Prime.

A single HD stream produces somewhere around 8 Mbits/sec. When all 3 channels of a single prime is running, that's approx. 24 Mbps; no sweat on a 100 Mbps switch.

You're probably still OK with running 2 Primes with all channels going simultaneously (48 Mbps). But I'd hesitate to recommend 3 fully-utilized Primes on a long-term basis. 72 Mpbs is getting uncomfortably close to the theoretical max on the 100Mbps uplink port.

Notice I refer to switches, not hubs. The above won't fly using an old 10/100 hub (too many problems with collisions).

Regardless of the bandwidth issues, I still think it's worth a try connecting the Primes via a 100M switch. At the very least, we'd learn whether the devices continue to drop off the network.

[KryptoNyte]

So based on this conversation, what would be the theory for why one of the Prime devices is quite solid while the other one constantly disconnects, both connected to the same gigabit switch?

[KryptoNyte]

Well, this is going to have to wait I guess as I have a few extra gigabit switches laying around, but no older 10/100 switches.

I still have an extra Ceton 4pcie tuner - I hate to revert to this, but it looks like I'll just pull the bad Prime and pop the cablecard in the ceton and then call for pairing.

Maybe Ceton devices will get popular again [and better supported] if SiliconDust can't get their act together.

[nyplayer]

Originally Posted by KryptoNyte

Well, this is going to have to wait I guess as I have a few extra gigabit switches laying around, but no older 10/100 switches.

I still have an extra Ceton 4pcie tuner - I hate to revert to this, but it looks like I'll just pull the bad Prime and pop the cablecard in the ceton and then call for pairing.

Maybe Ceton devices will get popular again [and better supported] if SiliconDust can't get their act together.

One thing have you replaced the Power Supply on the one that keeps dying ??

[KryptoNyte]

Originally Posted by nyplayer

One thing have you replaced the Power Supply on the one that keeps dying ??

That I have not done, I'll give that a shot, thanks.