First off, thanks for the emails you have sent me the last months, I love to hear about whats going on in your part of the world. That being said, the contact form is not a support line.. I fully understand how frustrating many of the ins & outs might be, hey, my blog is called rantville for a reason. If you come with relevant info to a thread, I’ll help out as best I can. Sharing the knowledge goes both way, and that would be lost in an email.. mkay?

With any MythTV setup utilizing more than a couple of tuners you are faced with a few things to consider. New motherboards rarely come with more than 2-3 PCI slots, and 1-2 PCI-E slots. Now, lets make matters worse, say you need to extend your raid array and are all out of PATA/SATA ports.. And.. we all know how rubbish onboard NIC’s can be. Unless you suck at math, or are so unfortunate that you have cut off half your fingers, you are all out of PCI slots by a good margin.

So, what do you do? Up till now, Ive been using somewhat older hardware, where PCI slots still are in majority. And, if I ran out, I setup a slave BE.. Fair enough, good heating, and the electricity company loves me(I better get a golden xmas card from them this year!).

Thats about to change now tho, Im going to try and become more ‘green’ with my computing this year. It wont be a major cut, but anything helps, no?

First step is to cut down on the number of servers thats running 24/7, thats currently 5;
2x OpenVZ servers(1x Quad ,1x Dual)
1x Fileserver(amd64 with a variation of PATA drives & a 4xSATA array)
2x Myth servers(one amd64x2 with 3xSATA array, other amd64 slave BE with nfs mounted array from Master)

The 2 OpenVZ servers got NFS mounted VZ storage from the fileserver, and run a vast variation of services. They are by no means utilized to its fullest extent, so this is where Im gonna make the first change.
The main servers(OpenVZ) are located in a different house on the farm than where my MythTV BE’s are. They will be split up in the coming revamp, Ive changed switches on all main exchange points to HP Procurve Gbit, this makes it all alot leaner.

I got ahold of a new(older) LGA775 mobo; Asus P5GC, which will be placed in the OpenVZ server with the Intel D940 CPU(Dual 3.2Ghz). All the tuners from the 2 Myth BE will be moved to this box, since it has 6 pci slots, I’ll also move over 2 of the SATA arrays(Chieftec backplane) to this maxi tower.

This will make both the old Myth BE redundant, and the old file-server will only hold the PATA drives thats done their duty well. I’ll shut it down when the majority of the drives have died, only right that they are given a proper retirement plan. They dont hold any data of value, all such is properly raided on new SATA drives.

The new MythTV BE currently got 3x DVB-S tuners(2xSkystar2 & Cinergy 1200-S) and one PVR-350, the latter is used with Zoneminder and the odd analogue captures I do of old VHS/other sources.

The hdd was allready setup with Hardy, and I see no reason to go with Intrepid here, Hardy got 2.6.24 OpenVZ kernel, thats not tagged as “stable” yet by the moderate OpenVZ devs. I had a few kinks with migrating VE’s from my Gentoo OpenVZ server(2.6.18), the ‘in-famous’ udev issues re-emerged. Ironic that it works flawlessly on Gentoo(Ubuntu VE), but not Ubuntu on Ubuntu? Easy fix, but I had to shake my head at that one..

8 hours in, and I got the best of all worlds, -2 servers sucking the wall socket, better utilization of the 24/7 servers. And still 2 PCI slots to spare in the new MythTV BE(ok, 1 really, Im still waiting for the pci sata card I ordered before xmas).

In case I dont post anything the coming week, I will be gone(from the 8th) till first week of february. So no respons from me before Im back, enjoy my absence. 😉

Ive spent some time doing research on the various aspects of this project. Read alot, asked around, pondered, and pondered some more.. What I do know without a shadow of a doubt, is that there are many unknowns, and quiet a few of those will remain such until they are tried out it seems.

Granted, it would have been a helluva lot easier if I had those deep pockets I briefly mentioned being absent in Part#1. So, Im left with trying to find those clever solutions to the issues at hand instead.
As a geek, that IS alot more fun, even tho it will take me longer to complete..

Let me try to draw you a picture and explain.. (up is north if there was any doubt)

hoiseth-sat-layout
(Like my drawing? Im good aint I? Riiight…)

Ok, what you see above is the farm I live at, the blue dots(#1-3) are where there is currently a sat-dish in place. #1 & #2 are 65cm dishes with the so called ‘Europe’ LNB setup; 1W, 5E, 13E & 19.2E. #3 is a old motorized analogue dish that will be replaced with either a ‘Nordic'(1W & 5E) or a ‘Europe’ dish as in #1& #2.

#4 is where we are currently planning to setup BigBen, ok, the name is not set in stone, Im open for suggestions.. The wall on that silo is around 5meters wide, we’ll either do so some serious steel hookup & bang it to the wall(2meter+ over ground), or place it on the ground with a concrete ‘flooring’. In this position it will be somewhat hidden from the recreational areas we use.

The alternative is to place it next to the garage south-west of mark #4. Not the best alternative for the following reasons;
– More digging, need to get the cable safetly underground between the silo(#4) & garage.
– Longer cable stretch(15-20meter)
– Visible from garden south of #3

Ive not done a precise measure of length, but its about 150meter from position #1 to the wall facing west on the silo(straight line). Adding about 20 meters going through cellar(#2) & loft in the barn is atleast a given, might even be more..
The red solid line between #1 & #2 is where we dug down a tube this summer for Cat5e, it has plenty of room for several cables.

If Im gonna do a straight shooter with the cables that means we’ll have 1 stretch going from #4 to #1(~170m), 2x from #4 to #2(~70m), and the last one from #4 to #3(~70m).

Fair enough? Well, somewhat..

In comes one of the big “unknowns”; How much signal & power loss will the longest cable suffer? After consulting people with the know-how, the results are not conclusive.
The signal loss could be within an acceptable %, if not I can put an amplifier at the middle(#2), but this will not help with the potential power drain. That is, the power the tuner/receiver sends to the LNB to switch between vertical & horizontal.
‘Bobben1’ at a .no board suggested I could overcome this with attaching a “power-supply” to one of the outputs on the LNB, granted a variable supply.

My question is this, could I instead use a multiswitch(at #2) using 2 of the outputs, and gain 4 outputs? With a Spaun SMS 341 F for instance?
This would overcome the need to switch between v/h..
With this solution(?) I wont loose 1 output, but instead gain 2.

This is important, not only for the loss of a output, but also since I have one MythTV box in service, and plan to setup a slave backend on location #2. The important bit is this; the tuners I currently use are SkyStar2 which only support DiSEqC 1.0, and that will fail miserably on a Multiswitch..
EDIT: Reading up on this, and I was wrong, there are several Multiswitches that can accomodate this need. For instance a EMP-Centauri E.143-AP, and hook one output to a 4/1 DiSEqC (with 1W & 5E fi), or directly to one of the SkyStar tuners.

The other tuners (DM500, DM7020 and soon DM600 on #3?) got no issues with DiSEqC 1.1/1.2.

Moving further, the existing ‘Europe’ dishes(#1 & #2) got 4/1 DiSEqC’s, either;

– Drop 13 or 19.2E input and screw on 28.2E
..or;

– Get a 8/1 DiSEqC to replace the 4/1 and keep all sat positions

On position #4 depending on what LNB setup we choose, either go for one of the above solutions, or if a ‘Nordic’ LNB setup is chosen, just use a 4/1 DiSEqC.. Pheew, easy!

The MythTV backend slave will be a dedicated 28.2E box with 1 DVB-S tuner to start off, replaced with a DVB-S2 tuner like the Net-UP when it becomes available. The Gbit link between houses #1 & #2 got more than enough bandwith to spare, so the slave(#2) will utilize the Raid array on the backend master thats living in house #1.

So to recap, I know more about potential show-stoppers on the technical side of the project. I still need to get more hands on experiences from people in the know.

Thanx to all the people thats taken time to respond to my questions & the suggestions you’ve given. Much appreciated! I do hope you keep it up.

Till next chapter..

New BBC channels on Canal Digital

| December 4th, 2008

I was looking forward to seeing the new BBC channels that Canal Digital started to beam yesterday. Granted, not alot of new content there, but still, more quality broadcasting is always welcome. More than enough crap being aired these days.

The channels are BBC Lifestyle, Knowledge, Entertainment & HD, yay!? well.. after tuning in last night, not so much. I dont have a HD tuner/receiver, so that channel is out of my reach atm, but the other three, well, take a wild guess what format they are aired in.. Thats right 4:3, completely disregard the original format, its actually aired in full screen. I could hardly believe my eyes..
Watching season 9 of Top Gear in a crippled 4:3 format is just too painful for my taste..

I hope this is just a “minor” premiere hick-up, but I fear it aint. For the person(s) making this decision I grant a well placed foot in their ass as a xmas present. Hohoho

Im a geek, and my fields of interrest range wide & deep. Ive had fasinations for many things over the years, some have stuck, others have faded. Faded for many reasons, some of them for financial reasons, lets face it, unless you have a deep pocket you cannot get all your heart desires. Such is life, no complaints tho, the feeling of finally getting something you have saved up for is just too grand for me to ever want to swap it for the “of I’ll just buy it, even tho I dont need it..”

Ever since I saw my first BUD, Ive been fasinated with what could be possible to bring home from the sky above. The first BUD I saw can barely be defined as a BUD, its was “merely” a 2meter dish, but since I was a youngster that was huuuuuge!

Ive dreamt of owning a BUD ever since then, but due to both finanical reasons and, well, living in city areas this has been impossible. The prices on a dish over 2meters is just mindblowing, its an easy 2 month salary..

I like to check out the used market, on all kinds of things really, be it music, instruments, movies, furniture, cars, computer hw.. anything goes, not that I buy that much, I just like to window shop.

Last week I was browsing around a .no market place; finn.no, and lo and behold, someone was selling a BUD, for the price of a regular 70cm dish! I sent him an email asking for details and if the price were right or if there was a zero missing there at the end. Price was right, so today I confirmed the purchase.

This will be my winter project(well, one of them anyway), it wont be setup before the winter is gone, but atleast the “biggest” investment is done. It will get a round of sanding & paint to “fit” into its surroundings.
Mark’s suggestion wont happen this time around, even tho I’d love to do it. 😉

The dream of a BUD is about to be realized, so is the urge to lock on to Astra2D on 28.2E, finally some quality tv!

I will document the process of transport, digging, setup etcetc in detail, for now, feast your eyes at my first BUD;

bud-4.8m-1bud-4.8m-2

I dont have the brand/name for it, but its a 4.8meter steel dish, complete with ground mounting & feedhorn. Whats missing of HW is a C120 LNB and cabling, I need some high quality coax for this setup since it will be placed about 120meter from the house. Gotta try and “hide” it away as best I can. (yep, some people dont see the beauty of such a dish.. go figure)

Invacom or Inverto ?
invacom-c120inverto-c120-quad

Ive had great experience with ‘regular’ Invacom Universal LNB’s, but I will see what prices the .no outlets can give me.

I welcome any feedback & experiences you might have with a BUD, would like to avoid the most obvious mistakes.

Till next chapter…

Mythbuntu 8.10

| November 4th, 2008

Another box, another update. This time around my MythTV backend(BE) was due for a upgrade. Been chunking along happily under Hardy for a good while, if it works, dont break it? No sir, not in my house!

Of course, with any change on an important service, backup of db & confs is a given…

When that was done, the upgrade itself went smoothly, the only hickup I had were with v4l-dvb. I pulled the latest & greatest to see what was going on with the new API(5) and its inclusion of S2API. It did compile, but did not play nice with open-sasc-ng, and after a little reading I saw patches where needed for both open-sasc & mythtv to make use of the new API.

If my Myth BE had been running on.. ohh say Gentoo, I’d give it a stab, but not on a pre-compiled distro. That would mean (re)compiling any coming updates to MythTV untill said patches had made it upstream..

So I took the easy way out, reverted v4l-dvb to revision 8984(pre API5), which works fine with latest open-sasc-ng & current Mythbuntu repo(incl 2.6.27 kernel). Due to the big stir up on multiproto vs s2api, Im putting a hold on a dvb-s2 purchase. Even tho s2api has been selected and merged into v4l-dvb, I’ll wait till the dust settles before I make a decision on what hw to get..

Since I run a headless BE, I cant speak much on any of the GUI changes, I did run mythbuntu-control-center via ssh tho, most notable changes there were prep’ing of images for diskless frontends. Good stuff for those that dont like to mock about in a terminal.

What I did notice, where an increase in overall performance/speed. Both mythweb, and all communication with frontends(FE). Channel lock & changes are cut with a few secs, this might also be due to open-sasc-ng, since I ran legacy code before upgrade, but had the same impression upon playback & editing of recordings. It all feels more snappy.

The speed bump made me want to clean up the channel list(long overdue), any change there via Mythweb in the past made the browser unresponsive for 4-5 minutes upon hitting ‘save’. Satellite hookups will do that I guess, 2500+ channels are a hefty load to process.
Cut it down to about 1/5th, which in turn made my BE grunt like a quad-core on speed.
This is still a part of MythTV Im not overly positive about, manageing a big channel list, with its xmltvid, commercial free or not etc settings. Be it mythtv-setup or mythweb, the last word that comes to mind is practical..

But, thats a MythTV issue, not Mythbuntu in particular, bottom line, Intrepid does a good job on my BE, my diskless FE’s, and heck, it even runs very well on my Aspire1 now that the Intel drivers for the GPU had a fix pre Intrepid launch.

UPDATE:
I have not dug into the amazingly long change log since 2.6.24, but to underline what I said earlier in the post(performance), have a look at the CPU usage & load on the box after the 2.6.27 upgrade..

cpu
load

There are no changes in the number of services on the box, nor has the recording schedules changed.

There was a constant load close to 2.00 under 2.6.24, this I filed under kernels inner demons cause I could not find a reason for it. Non of the apps or services justified this load, looks like they got to this demon. This is a Athlon 64 x2 4600+, with 2GB ram, SATA drives only..