The Little Woman Takes a Thin Client

Well, the Little Woman’s PC really became flaky and we looked around for alternatives:
“This Tutorial describes the downgrade process you need to run to get back stable after sid upgrade.”

  • a new PC
  • fixing the old one
  • a thin client which I have in a pile…

If she uses a thin client for a while she might like it and we can fix her old one at our leisure. Her flaky PC has a decent CPU but is low on RAM, which was expensive back when it was built. It might take only a new PSU and a couple of sticks of RAM to fix that baby up for a few more years of use. It has a beautiful case and a decent dual-core 64-bit CPU. Taking it off-line will also allow us to switch it to 64-bit software. She had 32-bit because OpenOffice.org wasn’t 64-bit back in the day.

So, I did the LTSP-thing in Debian Jessie and an old thin client worked beautifully, except when it didn’t. It took a while to get rid of an annoying GREEN screen (changed from via to vesa driver). Then it took forever to get X to load (fixed that by switching to openchrome… Choice is good. X now starts in a couple of seconds.). Unfortunately that still crashed occasionally. I was surprised that sound worked immediately. Systemd and pulseaudio did their thing on the thin client and mplayer on Beast spoke “pulse” just fine. But, it seemed that after any interruption, mplayer was denied further connection. It went from perfection to “connection denied”. So, at least two bugs plagued us with Jessie on the thin client and we decided rather than fighting bugs on what is supposed to be a production machine, we would “downgrade” to Wheezy. We know that works… Downgrade sounds so much like that other OS, eh? We are going from beta software to known good software. It works for me.

The downgrade was rather simple as TFA suggests. A few packages were broken. I anticipated that Systemd would break because it is so complex so I removed it before the downgrade. There was also trouble with ACPID, but I managed to fix that with aptitude. apt-get would not do the job. This was done on a chroot in Beast which gets fed to the thin client via PXE-booting with TFTP and NFS for the root file-system. I also downgraded the kernel from 3.16 to 3.2 in case the instability was with the newer kernel. I also obtained the code from Via for their driver just in case we needed to build our own for some reason. They also have archived binaries which match archived Linux kernels…
ls RedHat/7.3/c3
libddmpeg.so via_dri.so via_drv.o via.o via_v4l_drv.o videodev.o
file RedHat/7.3/c3/*
RedHat/7.3/c3/libddmpeg.so: ELF 32-bit LSB shared object, Intel 80386, version 1 (SYSV), dynamically linked, not stripped
RedHat/7.3/c3/via_dri.so: ELF 32-bit LSB shared object, Intel 80386, version 1 (SYSV), dynamically linked, not stripped
RedHat/7.3/c3/via_drv.o: ELF 32-bit LSB relocatable, Intel 80386, version 1 (SYSV), not stripped
RedHat/7.3/c3/via.o: ELF 32-bit LSB relocatable, Intel 80386, version 1 (SYSV), not stripped
RedHat/7.3/c3/via_v4l_drv.o: ELF 32-bit LSB relocatable, Intel 80386, version 1 (SYSV), not stripped
RedHat/7.3/c3/videodev.o: ELF 32-bit LSB relocatable, Intel 8
It’s all good.

One thing I noticed while testing this thin client was that loading applications onto the screen was snappy, like 2s, but changing tabs in the browser was very slow like 5s. I’m not sure what that’s about. Beast was not swapping at all. I reduced the screen resolution which made an improvement but it’s still poor. I’ll bet the instability and the slow performance are connected. The thin client was actually running out of memory even though it has 256MB and ordinarily only needed 64MB. It could be something silly like logging to RAM. I’ll figure that out. The audio was beautiful, unlike her previous PC, which had really poor sound. So, she should have 6 year-old performance from Beast on an 8 year-old small cheap computer in place of her 10 year-old PC. Good fun.

The downgraded thin client booted immediately and performed well except now pulseaudio doesn’t run. I get, after configuration tweaks to eliminate other messages,
“root@ltsp194:~# pulseaudio –system
N: [pulseaudio] main.c: Running in system mode, forcibly disabling exit idle time!
E: [pulseaudio] main.c: Daemon startup failed.”

I had a hard time to get it to run with --system, so I ran it per-user as root and allowing anonymous authentication…
“root@ltsp194:~# pulseaudio –start
W: [pulseaudio] main.c: This program is not intended to be run as root (unless –system is specified).
I: [pulseaudio] main.c: Daemon startup successful.”

Now, it complains and won’t start that way, but this configuration runs, exhibits all normal behaviour but emits no sound…
/usr/bin/pulseaudio --system --exit-idle-time=-1 --disable-shm --no-cpu-limit --resample-method=trivial --high-priority --log-target=syslog -L module-udev-detect -L module-native-protocol-tcp auth-anonymous=1 -L module-stream-restore -L module-rescue-streams -L module-native-protocol-unix auth-anonymous=1 -n

I was doing all this tweaking from Beast. When I went to her office, she was already there using it and she said it’s better than the old machine because the display fits better… ;-). She has a rather wide monitor. I eventually got pulseaudio to run with –system but it never made any sound from the speakers with the downgrade to Wheezy. I’ll give up on that.

Not bad for so little RAM:
root@ltsp194:~# free -m

total used free shared buffers cached
Mem: 216 125 91 0 0 28
-/+ buffers /cache: 97 91 119

The thin client and Beast are having a love-in. No crashes, glitches or flawed screen for days now. I just have to fix the configuration of pulseaudio and we’ll be good. It’s refusing connections and squawking about X… It’s just like Lennart to tie X and audio together… He doesn’t get the UNIX philosophy. I could tunnel the audio over SSH but it’s probably just as easy to fix the X-authentication. He was probably thinking there were multiple users and wanted X-authentication to sort them out. Sigh…

Got the X-authentication working but still no sound with Wheezy. No kidding. The player, pulseaudio and ALSA all are happy and still no sound… This is like peeling an onion. In the old days I would just fire up esound and it would work. Simpler is better for me. After returning to this matter, I found the thin client had dead audio-hardware. It died during my set-up just to maximize annoyance. I swapped another client and I had sound by ALSA but not by pulseaudio, and pulseaudio keeps crashing. Crap! I’m now looking at streaming audio from Beast to the client.

Well, the next thing to get working was printing. It turns out that’s not so easy. She was running i386 software because the printer driver is only available from Xerox in that architecture, and an rpm from old releases of Suse and RedHat to boot… I decomposed the rpm with rpm2cpio and set up a file-structure with cpio. Then I moved the files where they needed to go in a multi-architecture installation of cups:386. I set up CUPS for the printer and it worked right away. We can now print from any PC with cups-client installed via lp -d -h -U… I did have to set up /etc/cups/client.conf because otherwise some commandline parameters were ignored… Well, Beast is running beta-software.

I had to tune up xorg.conf on the thin-client’s chroot because DRI wasn’t working at first. It was pretty slow without (VGA compatible controller: VIA Technologies, Inc. VT8623 [Apollo CLE266] integrated CastleRock graphics (rev 03), really primitive by today’s standards). Now scrolling is pretty snappy although video covering more than 1/4 of the screen at 1024×768 is jerky and makes the system sluggish. She has TVs and other PCs for video anyway, so it’s no great loss. This thin client is for her work computing. Everything but sound is working as expected. I will fix the sound using esound, I guess, and move her files and the ftp server over on her hard drive which is OK. It will be trivial to upgrade her thin client. We could even fix up her old machine and swap it in minutes, just by undoing the xorg.conf stuff.

To sum up: While a couple of things have gone backward: video and sound, she now has the raw performance of Beast not swapping and using 500gB hard drives with 4 AMD64 cores. Compared to 2 AMD64 cores in 32-bit mode, it’s a rocket. Her old machine swapped a bit and was generally much slower although still usable. She’s just fine with the responsiveness of the thin client to the scroll-wheel or clicking the slider or punching Page-Down. Here’s how bad the network gets clogged with some screensavers. I turned those off… Some pages with Flash drive it nuts, though. I could just disable Flash… Network I/O: with her logged in but inactive, it’s taking min: 2.2 KB/s, avg: 4.7 KB/s, max: 10.6 KB/s in and min: 8.3 KB/s, avg: 16.6 KB/s, max: 25.6 KB/s out on eth0 with her active, courtesy darkstat. While browsing MrPogson.com, the input peaked at 170KB/s and the output peaked at 1.3MB/s, just a fraction of the network’s capacity. Maxing out the network connection with dd reached 10.7MB/s to the thin client. So, the bottleneck is the thin client. I made a comment and use of the page was snappy, just a slight hesitation to load a page.

Well, all this for one thin client is a bit much, but with gigabit/s networking and a room full of thin clients, it would be no more work. There are the unmistakable advantages that her thin client uses much less power and is the size of a box of chocolates…

See Downgrade from Debian SID to Stable from Jessie to Wheezy.

About Robert Pogson

I am a retired teacher in Canada. I taught in the subject areas where I have worked for almost forty years: maths, physics, chemistry and computers. I love hunting, fishing, picking berries and mushrooms, too.
This entry was posted in technology and tagged , , , , , . Bookmark the permalink.

42 Responses to The Little Woman Takes a Thin Client

  1. oiaohm says:

    DrLoser Lloyds of London is your biggest third party warranty/insurance company. If you buy a stack of cheap thin-clients the warranty cost will be high due to their shorter life span. Lloyds of London will charge you more than going out and buying HP clients on cheap clients warranties. This is where TCO gets tricky.

    Lloyds of London allows you to insure your operational equipment that it will be operational.

    DrLoser
    Because it only saves about $50 a year per desk, even when everything else works out.
    That figure is incorrect. http://www.devonit.com/blog/advantages-in-converting-to-desktop-virtualization

    The figure is somewhere between $100 to $200 dollars a year per seat when you do proper TCO including power usage and maintenance of switching over to thinclients. You don’t use the term per desk particular with thin clients as you might have 4 seats/users to a desk off a single thinclient device.

    2 to 4 times that 50 dollar figure. That 50 dollar figure is base on a very old and incorrect TCO assessment.

    Of course the 100 to 200 dollars can go completely out the window if the software you require cannot run. 80/20 rule applies to thin-clients just as much as VDI.

    Usage of thinclients needs to be based on assessment. Please note some places for 80 percent of their staff don’t have computer access. The remain 20 percent with computer access could in fact all be those who require full PC systems. Like customer being able to perform a price check at a terminal in the store that is a thin-client job but not every store will provide this.

    If your business does not have any thin-clients you really should do an assessment. Maybe you will find there are a few places where customer service or staff actions can be improved. Key thing about thin-clients is the fact they can be fixed to things. One example here was everyone working in warehouse was using tablets/phones lot of breakages. Thin-terminals connected to the lifts and carts damaged ceased and cases of lost mobile phones/tablets also ceased. If tablets and mobile phones in particular area are increase productively but costing a fortune it might be a case of the wrong tech in the wrong area. A full PC in those zones is normally not good either.

  2. oiaohm says:

    DrLoser
    Basically, and as long as you buy from a manufacturer who has long-term skin in the game (Hewlett-Packard being an obvious one), it doesn’t really matter whether the warranty is for one year or three years.
    It does matter it all about circuit board infant mortality. HP provides 3 years warranty because this proves they trust against infant mortality. Cheap electronic devices have 1 year warranties because they know the majority will die from circuit board issues between 2 to 3 years old.

    Ok from the time I first quote to now the t5125 has disappeared off the extended warranty. So it was a 13 year version.

    Oh, I thought thin clients were totally interchangeable. After all, they all run Linux.
    Today can be yes past absolutely no unless you made sure to get Linux ones.
    A t5135 happened come with HP Thinconnect OS . HP Thinpro is Linux or shipped as Linux 2.4 OS is Linux for sure. HP Thinconnect OS was vxworks first version all HP thinclients with HP ThinConnect first version were mega lemons due to software problems at least it was fixable by installing ThinPro that HP wanted to change for.

    HP Thinconnect first version did not work with HP software for central management (Altiris Solutions) Hp in the time of t5135 release expect you to pay extra to have HP Thinpro the Linux version. Basically clients 1 were disappointed with t5135 and t5145 two absolutely rope-able that HP had the idea they had to pay extra per thin client for central management on top of paying so much per machine for Altiris.

    Hp had vxwork love affair on thin clients that lasted about 2 years. All those vxworks beasts have an upgrade to HP Thinpro for free these days.

    Altiris configures pure Linux thin clients back in the day today it does Windows embedded clients as well.

    DrLoser a Linux thin terminal is good. So strange OS from hell thin terminal prepare for hell(like vxworks based ones and the like we almost never find any more). Windows OS thin terminal expect to run out of date software.

    DrLoser googing without understanding the topic of thin-clients is going to lead you into error after error. Yes first releases t5135 suxed with it default software. Yes it really suxed that you had to individually update each one. Why HP Thinpro you have central management of updates. All HP Thinconnect OS thin clients you will find some reports about upset users.

    Please note some sites will tell you that t5135 had Linux HP Thinconnect. HP in the end reimaged a stack of the t5135 before they went out door with HP ThinPro re-branded to Thinconnect OS. Lets pretend we did not make this error all the re-branded connect to Altiris Solutions. So if a t5135 will not connect to Altiris you have something to disinfect quickly.

    There are a lot of complaints around ThinClient t5135 and t5145 about connecting to windows servers these normally turn out people have not updated the software on their thinclients and have originals.

    By the way you can boot a usb key in both t5135 and t5145. Yes they run debian and thinstation.

    DrLoser there is a reason never to be first adopter.

    http://www.parkytowers.me.uk/thin/hp/t5730/mods.shtml
    What users sometimes do to thin-clients is turn them into stand alone machines today due to how powerful thin-clients are getting. Linux users don’t just convert Windows PCs into their computers.

    So expanding thin-client sales might not mean thin client. ChromeOS on desktop can be done in a lot of current day thinclients.

    HP ThinPro OS is based on Debian.
    http://h30499.www3.hp.com/t5/Workstations-Thin-Clients/ThinPro-3-3-Dual-Monitor-RDP-Working/td-p/5497223
    So much that you can in fact use Debian repositories and Debian kernels. HP is not the only one shipping Debian based thinclients.

    This is a problem. Claiming that Linux is not on shelf in a store can be kinda wrong. The store has thin-clients, Debian can truly be on the shelf.

    DrLoser do you get it now Debian does have deals with OEMs. Problem is none of the deals put stuff in front of people that they would see as a normal PC.

    Current day HP thinclients there is absolutely nothing stopping them being converted to Debian machines.

  3. DrLoser says:

    And yet, as always, I endeavour to be helpful. And particularly around the festive and gift-giving time of year.

    Have you considered splashing out 199 Loonies on a Chromebook, Robert?

    I am not acquainted with the Little Woman, but I imagine she would appreciate the (very much within budget) gesture.

  4. DrLoser says:

    Apologies, let’s try again. Now, the point of your post, Robert, I believe was this:

    Her flaky PC has a decent CPU but is low on RAM, which was expensive back when it was built.

    And is a sunk cost. Any new thin client, no matter how ridiculously cheap (unless you sign a personal contract to ferry a single item from AliBaba), is going to have 2GB (bottom end “Chromebooks”) or 4GB RAM.

    It might take only a new PSU…

    Dear me, no experimentation? I’m sure you’ve got piles of unused PSUs sitting out back there.

    and a couple of sticks of RAM to fix that baby up for a few more years of use.

    I imagine there might be other minor issues, which I will leave to hardware engineers and their admirable ilk to point out.

    “A couple of sticks of RAM?”

    It ain’t no Joe Bazooka bubblegum any longer, ya know.

  5. DrLoser says:

    Her flaky PC has a decent CPU but is low on RAM, which was expensive back when it was built.

    It might take only a new PSU and a couple of sticks of RAM to fix that baby up for a few more years of use.

  6. DrLoser says:

    Many organizations find that about 80% of use-cases fit with thin client: data on the server where it belongs, no video, application on the server where it belongs, and that’s it.

    More hand-waving. Numbers, please, not “many.” And while you’re at it, could you partition the numbers between what I will call professional thin clients … ie big ole monitors in places like Lloyds of London … and cheapskate domestic thin clients, of which your own domicile would be a perfect example. (And I’m fairly sure you save a hundred bucks or so a year, because you know what you’re doing, and you have an excess of spare time.)

    It isn’t really a matter of whether you or I are correct here, Robert. That is merely willy-waggling.

    Fact is, outside edge cases like massive data-clerk systems, nobody out there is buying this Thin Client Schtick. Wanna know why?

    Because it only saves about $50 a year per desk, even when everything else works out.

    No management in the Western World would fart around with $50 per desk, unless there was a tremendous amount of desks (as in large insurance companies).

    Do you really see a future in “pure” thin clients that require a monitor behind them?

    Or do you see the future as a Chromebook, possibly with a docking station and a hot-seat monitor behind it?

    I mean, I’m trying to think about this stuff, Robert. As far as I can see, all you’re doing is to be an ill-considered reactionary.

  7. DrLoser wrote, “I’m struggling to think of a use case for an item that you can’t pack in your bag and take home”

    My thin clients have a Kensington Lock thingy so some employers may not want folks to take their machines home. It depends on the business and there’s no reason at all an employer could not have a mix of thin clients and others to meet every need. Many organizations find that about 80% of use-cases fit with thin client: data on the server where it belongs, no video, application on the server where it belongs, and that’s it. That would cover most clerical jobs, folks browsing/typing. Banks love them for tellers, for instance.

  8. DrLoser says:

    DrLoser, living in never-never-land, wrote …

    Happy Christmas, Robert!

    I’ve just been told by my (IT) employers in “never-never-land” that they’d already taken off the three working days between Christmas and the New Year from my holiday allotment, so I’ve an unexpectedly large amount of free time here in “never-never-land” to contemplate such things as Randomly Plucked Statistics … which appear to be a staple of your world. I will label that world “Provincial Retirement.”

    Or, if you prefer, “Everything I needed to know, I learned after I didn’t have to convince people that I was right to make a living.” Not very snappy, I know, but it will have to do. So, to your cite:

    IDC expects the market to record single-digit growth again in Q4 2014 and then decline marginally in 2015 by about -0.2% year on year as a result of a decline in the Central and Eastern Europe, Middle East and Africa region (CEMA).

    Wowza!

    And you’re looking at roughly 100K thin client shipments per quarter over the whole of EMEA.

    Wowza!

    Look, I’m not arguing that thin clients are insignificant. You’ll be pleased to know that I originally thought that they were, but this site has persuaded me otherwise.

    Apparently there is a huge number of noxious massive MegaCorps in EMEA, and presumably in the USA and Canada also, who force these crappy solutions down the throats of their demoralised employees.

    But, hey, I am not one of those employees. Putting myself in their place: maybe they watch a retro drama like “Mad Men” and long for the opportunity to sit in the modern equivalent of the typing pool and be groped by their bosses?

    I’m just trying to make sense of the trends, in a totally OS neutral way.

    Obviously, there are some very large entities that are choosing thin clients. (In my opinion they are regressing, but my opinion does not matter.) I’m trying to identify the use cases in question, because I’m afraid that just shaving $20 off your yearly hardware bill doesn’t cut the mustard, even at a Berkshire Hathaway company — I work for one, with a $3000 laptop.

    Nope. Ironically, oiaohm has led me to this tentative conclusion, and for that I thank him.

    If you want a massive monitor for things like Bloomberg, and all you need is a couple of 100GB ports to deal with co-located data from a couple of financial feeds, then Thin Clients are good. And, incidentally, the “warranty” element of them is close to useless.

    For anything else? I’m struggling to think of a use case for an item that you can’t pack in your bag and take home … eg a Chromebook.

  9. DrLoser says:

    A concession to Robert’s valid point here:

    I’d bet the price increment would be small because after 5 years the thing is burnt in and all the infant mortality is done.

    Whilst I would disagree about the on-going longevity of such a thing (if we’re going to anthropomorphise a thin client and assign it an infant mortality rate, I think we owe it to the poor dear to assign it an equivalent period of senescence, eg busted caps and denatured thermal paste), I think this is an excellent point.

    Basically, and as long as you buy from a manufacturer who has long-term skin in the game (Hewlett-Packard being an obvious one), it doesn’t really matter whether the warranty is for one year or three years.

    Why? You’re very unlikely indeed to need it for years two and three.

  10. DrLoser, living in never-never-land, wrote, “I would imagine that they have figured out a per-desk sweet spot for an utter piece of diseased useless computing crap attached to a very expensive monitor.
    This is not an option available to the world at large.”

    Meanwhile the percentage of PCs shipped as thin clients keeps increasing.

    Do the maths. Other people can as well. Tally up the total cost of legacy PCs idling all over the businesses versus thin clients idling and servers working hard and anyone can save a bundle. It does make some effort to install thin clients but the break-even fleet is just a few machines. Certainly it makes sense in computer labs in schools.

    e.g. a bunch of thick clients each with hard drive and 2gB RAM compared to a bunch of thin clients with minimal RAM and a server with 50MHz of CPU-cores per user + 512MB RAM + 100 MB per user. Just multiply the difference by N and count your savings. Systems that I built cost ~$25 per user on the server + $100 thin client and $100 monitor+keybaord+mouse back in the day are probably wasteful compared to today’s price/performance. It’s somewhat like the return on investment of cloud-computing. Put the computing power and resources in the server room more efficiently than replicating them all over the network. e.g. I was in computer labs with 512MB per PC with 24 machines in the room producing 4KW heating. We could easily run that lab with a few gB on the server and just a few hundred watts of heating. Do the maths. Thin client computing saves a bundle, especially with $0 for licensing with GNU/Linux. Performance, except for full-screen video was superior in every way for students and the applications they ran: word-processing and browsing, mostly waiting on the user or the web. They never had to wait on storage because all their applications were cached in RAM.

  11. DrLoser says:

    Just to remind ourselves of oiaohm’s customary First Absurd Position, which he has now abandoned without so much as a tiny squawk of apology:

    Reason they have 10 year of hardware testing when they are new.

    Reason this went away with the first dawning of the light. It was presumably a back-brain response to my mention of lead acid battery testing.

    It’s really easy to twit oiaohm if you know the right buttons to press.

    A thin-client is about the only form of computer you can get a 10 year warranty on.

    Which has now, mysteriously, dropped to a 5-8 year warranty. Without explanation. Why? Because oiaohm cannot google up a convenient link; presumably, such a link is lacking.

    The long warranties and low price is why thin clients can be so cost effective.

    This is nonsensical, even given all the assumptions.

    Let’s say I can get a Thin Client at $250, which is not unreasonable if you’re thinking about a Chromebook or equivalent. Now, either the warranty is built into that cost, or it isn’t. You know what?

    I’d far rather it wasn’t.

    If I buy fifty of them (and who the hell buys three or four thin clients at a time? What sort of loonie would even countenance that?), then I’d rather ignore the warranty and just admit that, over (say) five years, 60% of them are going to die and are going to need replacing.

    At a lower cost, for better hardware.

    This “warranty” drivel? Even if it were accurate — and nothing oiaohm has come up with has suggested that it is — it’s an abnormal fantasy for practically any business use I can think of, outside massive government departments and insurance businesses.

    And even for those groups, it’s frankly of marginal benefit.

  12. DrLoser says:

    8 to 15 years warranties on thin-clients is not abnormal. You have to pay for it.

    I agree. (Although I would like either a cite or a quantification for “not abnormal.”)

    You have to pay for it.

    Let’s take any 50-500 employee SME you care to think of. How much do you have to pay? What is the cost/benefit analysis?

    Max warranties at min cost require knowing what you are doing with thin clients.

    Oh, I thought thin clients were totally interchangeable. After all, they all run Linux.

    And now you’re claiming that there is some way to get a maximum length warranty at minimum cost that involves something more than, er, comparing the number of years in one column against the amount in dollars in the other?

    Wow! Aren’t Thin Clients magical!

    Alternatively, this might just be magical thinking by oiaohm.

  13. DrLoser says:

    HP Compaq T5125 yes the one I first quoted

    You may have fat-fingered that one, oiaohm. It looks like what you really found was a T5135, which is apparently the next generation on from a T5125. And, interestingly enough, apparently a disappointing replacement.

    Released circa late 2006 or early 2007, I believe.

    DrLoser I can still buy another 5 years warranty on that sucker even that its 8-9 years old.

    Or 6-7 years old. And a replacement for the actual thin client that you described.

    And there’s a huge difference between getting a ten to fifteen year warranty on a product up-front, and a five year (hardware only, mind you) warranty on a product that’s, to generously take your terms, 8-9 years old.

    I’m going to let you guess what that huge difference might be.

    Now, leaving all that to one side. Let’s discount the cost of fluff-filled mice and coffee-damaged keyboards and various other minor inconveniences. What do you do with one of these wondrous Thin Clients?

    Why, you attach a monitor to it, of course!

    Robert generally leaves this out of his calculations, and not unreasonably, because to Robert, a “thin client” is a re-purposed XP desktop or the like.

    But in the real, commercial, world, a thin client comes with a monitor. And to take your example of Lloyds of London, oiaohm, quite a massive big expensive one.

    I’m not privy to the internal accountancy reports of Lloyds of London, but I would imagine that they have figured out a per-desk sweet spot for an utter piece of diseased useless computing crap attached to a very expensive monitor.

    This is not an option available to the world at large.

  14. oiaohm says:

    The HP care warranty is a suxer trap. Thin client comes with 3 years out box. Date of purchase of the care warranties is when it starts ticking. You will be able to purchase care pack for sure while the define has shipped warranty.

    Correctly purchased all HP thin-clients come with 8 years Warranty. Incorrectly purchased HP thin-clients come with 5 year Warranties for the same money. If the thin client turns out to be a good solid model you will get 13 to 15 years of care warranties out of HP before you cannot 100 cannot purchase another care pack for it.

    8 to 15 years warranties on thin-clients is not abnormal. You have to pay for it. Max warranties at min cost require knowing what you are doing with thin clients.

    Only people who deal in thin clients would be aware of items like care packs that stack that come from many thin client vendors.

    HP is one of the more expensive thin clients. The expense is compensated buy the length of possible warranty.

  15. oiaohm says:

    HP Compaq T5125 yes the one I first quoted http://h30094.www3.hp.com/product/sku/10446917#nav71

    DrLoser I can still buy another 5 years warranty on that sucker even that its 8-9 years old. Apparently its going to have a 15-20 year life span by HP. I am not exactly sure I really want to run a thin-client for 15 years. Getting monitors become a serous pain in but.

    And yet you can’t cite even one
    I cited one right of the bat with a 10 year warranty yet you were too dumb to know it DrLoser.

    Lloyds of London is who you use if you want a 10-15 year thin client warranty up front.

    HP does not say Linux when you read thinclient specs “HP ThinPro” and “HP Smart Zero Technology” are HP Linux Distributions for thin clients.

  16. DrLoser says:

    And yet you can’t cite even one.

    Immediate retraction: you can cite one. I’d be interested to see the small print.

    But is this common? And let’s face it, five years is still less than ten years.

    And to be quite honest, I’d happily rely on a Chromebook for five years, give or take 18 months or so.

    What possible benefit could I get out of, say, an Igel UD10-W7 Touch at £843.70?

    To be fair, the same model without “touch” is a hundred quid cheaper at £743.70.

    I think it’s fair to say, however, that Igel are not really aiming at the consumer market here.

  17. DrLoser says:

    Many components of thin clients have >300Kh MTBF, and 10 years is, what, 90K hours?

    It only takes one component to fail, Robert.

  18. DrLoser says:

    5 year warranties are highly common on thin clients. Some are strange. Like igel you register you company as using their brand of thin client and then every one of their thin clients you buy automatically gets a 5 year warranty.

    And yet you can’t cite even one. Not to mention that a five year warranty is a substantially less useful warranty than a ten year warranty … which is what you fantasised about in the first place, oiaohm.

    Good god. These things don’t even degrade in a linear fashion, pretty much by definition.

    Oh, and Igel thin clients? Here’s an End Of Life Extra Special Xmas Promotion for you. Who knows when this wonderful thin client first came to see the light of day?

    Well, perhaps you do, oiaohm. But I’ll bet you’re not going to tell us.

    And bear in mind that this was your cite. I have nothing to prove. You do.

  19. oiaohm says:

    https://www.igel.com/company/legal-documents/product-warranty.html
    5 year warranties are highly common on thin clients. Some are strange. Like igel you register you company as using their brand of thin client and then every one of their thin clients you buy automatically gets a 5 year warranty.

    Many components of thin clients have >300Kh MTBF, and 10 years is, what, 90K hours?
    Robert Pogson this is so correct it is not funny.

    To be correct is 5 year warranties for replacement with the same model in some cases. 10 years warranties are the last 5 years of warranties replace with equivalent current model. Lloyds of London is who you looking for a specialist in third party warranties.

    You have hardware makers warranties and then third party warranties. The longest warranties you can get in thin clients is 10 years without insane cost. Its less than 10 dollars per thin client. Please not its not every thin client can you get a 10 year warranties on. It all comes down to the components the thin client is made from having large enough Kh MTBF that Lloyds does not believe they will ever have to pay out. In fact more often than not Lloyds is right. Lloyds projects thin client deaths at about 15 years old for good quality high life span parts constructed thinclients.

    Dr Loser HP Compaq T5125 are second hand on amazon and ebay and many other places with some parties insanely asking 200 USD for them even that they are all nearly 10 years old. That model was only manufactured for 12 months and that is common

    I did not say 10 year manufacture warranties. 15 year contract is a max thin client life contract that Lloyds will warranty for. Costs on approve clients to 5 to 10 years is flat 1 fee. For 11-15 years the cost increases each year. When Lloyds thinks it will fail they increase what you have to pay.

    Reason why you see some manufactures doing only 12 month warranties on some thin clients is not that there operation life is short its a simple reality that the model is build from end of line parts and will be manufactured for less than 12 months. Most of the 5 year manufacture warranties have a swap with equal model after 12 months because for the simple reality there will be no inventory left of that model and no more parts to build any of that model ever again.

    10 years old on Linux thin clients when you start running into software support problems.

  20. oiaohm quoted oiaohm, “Some thin clients come with a 10 year warranty”, and objected…

    I haven’t heard of any of those but it surely would be feasible. Just charge a high enough price to cover the possibility and keep a reserve stock of replacements. Ask the accountants to give you a number.

    Here’s one with a 3-year warranty. Here’s another distributed by a Uni, “5-Year Warranty, 7-year expected working life”I’d bet an engineer could tell you the failure rate right up to 10 years and sell a warranty for it. I’d bet the price increment would be small because after 5 years the thing is burnt in and all the infant mortality is done. Many components of thin clients have >300Kh MTBF, and 10 years is, what, 90K hours? How many parts do the modern thin clients have? 5 chips (CPU/graphics, RAM, PSU, etc.), a motherboard and some connectors? That’s about right. How long does a piece of silicon last at room temperature? I’d bet someone would sell you a “maintenance subscription” for a certain amount, provided your secretary kept a few in stock to swap… That’s not much different than paying for insurance/warranty.

    Here’s a 15-year contract guaranteeing energy-savings with the use of thin clients of $64million. The entire cost is less than $0 because of those energy-savings. Clearly, those folks aren’t worried about the devices having to be replaced very often… but they could have saved much more by avoiding use of that other OS.

  21. DrLoser, quoting oiaohm, wrote, “thin clients are obsolescent equipment when new”.

    I think it’s true that thin clients tend not to be state of the art PCs. There’s just no need for that kind of power unless you do a lot of multimedia and can’t run the data over the network. That’s rare. With gigabit/s connections, full-screen video is easy. It’s the server that does most of the work. Consider the legacy PC. I bought one in ~1992 with 1024×768 graphics. With that other OS it ran 800X600 for years. It finally died in 2002 when I shipped it with inadequate crash-resistance up north. 1024×768 is still widely used although it’s on its last legs. So, a thin client set up ~1992 would be just ending life now if all it had to do was show the pix and send the clicks. That’s 20+ years. Typically, folks keep them ten years and change them for one reason or another (fanlessness, smaller size, higher screen resolution, different video-standard, resolution, more RAM, dust-bunnies or faster networking connection). With many monitors now happy with much higher resolution, folks are motivated to upgrade them but, for many purposes, there is no need, just like telephones don’t get swapped very often. The Little Woman’s thin client is obviously obsolete but it still does the job and likely will be kept a few more years. Except for full-screen video it has no problem at all doing the job. We have TV for that. We have several…

  22. Dr Loser says:

    Incidentally, most thin clients I can find at the moment have at most a 3 year warranty and quite often a warranty as low as 1 year.

    I won’t trouble you with specific links. Just google “thin client warranty,” which is what I just did.

    Ten years?

    You seriously owe us an explanation for this outrageously unlikely claim, oiaohm.

  23. Dr Loser says:

    You would be referring to one of these. HP Compaq T5125 is a good example of a 2005 Linux thin client.

    I wouldn’t be referring to any such thing, oiaohm. But let’s say I’m gagging for one as a replacement for my Thin Client Farm.

    Do you, perchance, know where I could get a second-hand one?

  24. Dr Loser says:

    DrLoser thin clients are obsolescent equipment when new. This is how come know unless you got unlucky that you have 10 years of operation without issue with a thin-client because all parts that build a thin client have had prior life for more than 10 years.

    “Obsolescent equipment when new?”

    Not the finest piece of marketing I can think of. Then again, the run-on sentence was High Quality Frontier Gibberish, wasn’t it?

  25. Dr Loser says:

    Reason they have 10 year of hardware testing when they are new. A thin-client is about the only form of computer you can get a 10 year warranty on. The long warranties and low price is why thin clients can be so cost effective.

    No doubt you have some sort of relevant cite for this “ten year or hardware testing when they are new” thing, oiaohm?

    And may I trouble you for this “10 year warranty” thing? A cite, please.

  26. DrLoser wrote, ” 2005. Let’s guess at a price of $400. Let’s even allow it the unlikely eventuality of being Linux based.”

    There were a lot of GNU/Linux thin clients in those days. This one cost a bit over $100 + freight to the Arctic… Thin clinets on the low end are not intended to do much video. The processor is too slow at higher resolutions. This one can do ~2fps @1920 but 60 fps @1024×768. That’s good enough for most browsing. We have wide-screen TVs that do YouTube if we need more. This thing is for pointing, clicking and gawking. It will be useful until it dies. It can even serve as HTPC/CUPS/DNS server. I bought 10 of them for a presentation at a conference. Winter and food-poisoning kind of wrecked that but I’ve used them in several schools where I worked. While they come with GNU/Linux in Flash, I boot them PXE to have total control so it runs the latest GNU/Linux. At the moment it runs Jackd, X, ALSA and openSSH-server and it idles most of the time Any kind of full-screen video or even a bit of FLASH maxes out the CPU. 1/4 screen Youtube is OK. I think these units will outlive FLASH.

  27. oiaohm says:

    DrLoser this is a case you don’t know the topic. Most new thin clients are built from silicon designs that are at least 8 year old the average is 11 to 12 years old.

    Let’s say you bought your thin client back in 2005. Let’s guess at a price of $400. Let’s even allow it the unlikely eventuality of being Linux based.
    http://h10032.www1.hp.com/ctg/Manual/c00530687.pdf
    You would be referring to one of these. HP Compaq T5125 is a good example of a 2005 Linux thin client. CPU and motherboard chips in it is already been in production for over 10 years. Yes the CPU and motherboard chips started production 1994 so fairly much debugged by this point. Yes it clock speed is only 400Mhz way slower than a common 2005 computer. The clockspeed is the only thing that has been ramped up.

    That Linux (presumably embedded) would either be bleeding-edge 2.6 or, more likely, 2.4. It’s not going to be amenable to updates: it’s embedded.
    Wrong on the updatability Linux based thinclients from 2002 on were designed around update-able flash. HP Compaq T5125 and most of the other in fact are support by Linux 3.x kernel. This is Linux insane hardware compatibility.

    Windows base thin-clients from any time frame are trouble to update. If you are after long life with updates out your thin-client buy Linux ones.

    God knows what the hardware support for video is going to be like, but it ain’t going to be pretty.
    Other than been analog vga perfectly fine they did include H.264 and other common codeces back then. Its only if we start seeing lot of H.265 will there be problems. Go read the 2005 spec sheet there is no requirement to guess.

    “1280 x 1024 60, 75, or 85 Hz 8-, 16-, or 24-bit”
    This is fairly much standard for 2005 thin clients. Or in other worst standard HDTV result-ion.

    Oh, and also the screen is probably on the fritz by now.
    This is you only major issue. having to replace the screen.

    DrLoser thin clients are obsolescent equipment when new. This is how come know unless you got unlucky that you have 10 years of operation without issue with a thin-client because all parts that build a thin client have had prior life for more than 10 years. About the only thing modern in a thin client is the accelerated video play back if you use the vesa driver that is disabled.

    Thin clients is normal the last item new you will find a cpu in before its obsoleted forever. Its sucking the last dollars out a production line.

    Most Linux thin clients don’t include ACPI. In fact chromebooks don’t include ACPI. Linux kernel supports direct power management control does not need bios crap. Result is it easer to run Linux on a thin client than a standard desktop computer.

    DrLoser I know this is most likely complete strange that there is something IT that has 10 years of testing behind it. Thin client are a abomination in more ways than 1. Some thin clients come with a 10 year warranty. Yes new thin-clients are still use 10 year old silicon designs. You call the thin client market highly consecutive.

    Now, that $400 is a sunk cost. It has already been amortized to nothing.
    Remember by tax the amortized rate has to be expected life span and this cannot be less than warranty. So many of those $400 thin clients are still on deprecation sheets worth 40 dollars. They would be amortized next year or the year after.

    DrLoser this is another case of you commenting on a topic you have absolutely no knowledge of. So you loved insulting me when ever I do this. Really you are only a Loser. You are presume that everything I say is wrong. I have tricked so perfectly its not funny. DrLoser you are just a troll are non troll would have done some research into why thin clients have a operation life of 10 years. Reason they have 10 year of hardware testing when they are new. A thin-client is about the only form of computer you can get a 10 year warranty on. The long warranties and low price is why thin clients can be so cost effective.

  28. DrLoser says:

    DrLoser thin clients have annoying operational life of 10 years perfectly to about 15-20 years with some defects.

    This is an interesting claim.

    It’s particularly interesting to me, because I started my career working with lead acid batteries (for a Canadian company, no less). Every day I would walk past an array of test rigs designed to load-stress various battery designs over a projected lifetime. Some of these test rigs had been going for seven years or more.

    Now, clearly, this is the only way to make a quantitative judgement on the matter. And, slightly less clearly, it’s a completely pointless thing to do.

    Why? Because, by the time you’ve finished the longer test cycles, you’re already dealing with obsolescent equipment; something that, basically, there is no longer a market for. It seemed like a great idea in the 1980s, when tech cycles were far longer, but it’s a remarkably silly thing to do these days.

    Let’s say you bought your thin client back in 2005. Let’s guess at a price of $400. Let’s even allow it the unlikely eventuality of being Linux based.

    Now, that $400 is a sunk cost. It has already been amortized to nothing.

    That Linux (presumably embedded) would either be bleeding-edge 2.6 or, more likely, 2.4. It’s not going to be amenable to updates: it’s embedded. Anything like modern ACPI is probably unachievable. God knows what the hardware support for video is going to be like, but it ain’t going to be pretty. Oh, and also the screen is probably on the fritz by now.

    Look, there’s nothing wrong with the idea of a thin client per se. It’s a personal choice. Almost nobody on earth makes that personal choice, but it’s still a personal choice.

    But if you’re going to make that personal choice, I would suggest that stumping up $200 every five years or so for a Chromebook (today) or a FantasmaGizmo (in five years) is the way to go. Only seriously warped misers try to eke out the last bit of service from hardware that would belong only in a museum, except that it’s so broken and inherently uninteresting that no museum would want it.

  29. oiaohm says:

    Robert Pogson
    Speakers usually require their own amplifier so you can’t charge that to the audio-chip.
    Please note the word usually. Some like lots of the VIA chips are designed to directly drive 8 Ohm speakers. Compact space thin client with embedded speakers based on VIA chips do exist without extra amps.

    What’s the power-consumption of the filter?
    Problem here is not that simple what is eating the watts is not just a filter. If it was just a PCM and a filter you would be right there would be bugger all difference between buffers filled with zeros and no buffers. Problem is audio devices are not that simple.

    optional features such as tone and loudness control, and 3D stereo enhancement.
    From the VIA spec of the item you are handling. Notice the 3D stereo enhancement with tone and loudness control. 3D stereo enhancement control is part of an Audio Processing units. Yes you have GPU for graphic processing units and you have Audio Processing Units. Wrong actions will keep the Audio Processing units awake when it could otherwise be powered down. How you are keeping the Audio Processing Units awake will define how much power it eating. Like 6 channels out and its doing 3D stereo enhancement on zero filled buffers. Hello wasting power in a fairly large way. How does the Audio processing unit know it has nothing todo is by having no buffers.

    Audio is not just about making the sound come out the speakers it all the processing before it gets to the PCM.

    Robert Pogson “turning off amps and other things. ” The other things list in a audio output device can be quite a long list. Amps, Audio processing units, MIDI, Effects generators.(yes some have some analog effects that are not officially part of the ), buffering ram and any other extra audio design designer decide to throw in.

    0.5 to 3watts 3 watts is quite a complex audio device via does do some fairly complex APU items.

    Robert Pogson basically audio is a simple little so and so that everyone thinks is super simple. Treating it as a super simple when it a simple equals bad power usage. Correct process to handle audio devices so they can go to sleep when they should is not hard. Yes it simple don’t keep buffers around connected to the audio device when you don’t need them. Zeroing buffer out might sound the same to user but that tells the APU to stay awake.

    Sorry Robert you have over simplified how audio generation in computers work. Yes its fairly simple.
    Buffer from OS goes to the audio digital processing and buffer management then to the PCM then to smoothing filter then possibly into a powerful amp.

    Yes you want the audio digital processing to go to sleep when there is nothing for the user to hear. The kernel does not need to send any extra form of signal other than end all buffers. No buffers audio digital processing sections of silicon can turn off.

  30. oiaohm wrote, “Audio annoying eats more power than network or video out. Power usage of the audio can in fact exceed the CPU. Audio is expensive.”

    Crap! Ever heard of PCM? That’s Pulsewidth Coded Modulation. It’s digital, on or off, a duty-cycle run through a filter to generate the output to speakers/headphones. Digital generates heat mostly at the transitions, a very low duty-cycle, and it’s one switch per channel, a very low parts-count. Even audio amplifiers can be run that way. What’s the power-consumption of the filter? Almost nothing for headphones, because they are low power and high impedance. Speakers usually require their own amplifier so you can’t charge that to the audio-chip.

  31. oiaohm says:

    Robert Pogson Via Samuel 2 400MHz CPU and the VT8235 “South Bridge” bridge system around them keep there power usage down by effectively turning off areas of silicon they don’t need. This is not OS power management.

    Yes particular have huge stacks of hardware power management. 3w and 10w power usage is a lot of hardware design work. Issue is particular software will dig out glitches in hardware power management. Pulseaudio is one of them.

    Audio on via hardware can be consuming 0.5 to 3 watts to power the audio outputs. So yes it quite a decent chunk. Audio annoying eats more power than network or video out. Power usage of the audio can in fact exceed the CPU. Audio is expensive.

  32. oiaohm wrote, “Audio device power management logic is quite simple. If it has no streams open it can shutdown turning off amps and other things. Yes it take less than millsecond for audio device to bring its amps and other things back on-line. Audio devices sleep that often that their power management is not normally managed by the OS kernel but are hardware features.”

    I doubt that’s much of an issue on my 10W thin client boxes. It’a Via Samuel 2 400MHz CPU with 256MB RAM and 100 mbits/s NIC and Chrome video. The CPU draws only 3W full out. The RAM and Video are probably a bit more than that so the box just gets a bit warm with no fan at all.

  33. oiaohm says:

    Robert Pogson
    pulseaudio runs sound through ALSA which is not aware of power at all. That’s all controlled by the kernel which knows nothing of pulseaudio.
    This is not understanding. The some of the kernel actions of power management is based on what application applications are doing. Audio is one of those things.

    Why ALSA device has buffers connected it cannot be put into short term sleep by the power management in the hardware.

    ALSA itself contains no power management controls. How you operate ALSA decides if your program allows hardware power management to manage an audio device power effectively. To run ALSA properly for power management effectiveness includes disconnecting when you have no audio to send and only connect to the audio device when you really do have audio to play or record.

    VT8235 “South Bridge” yes the audio in those does power down when it has no connected audio clients. This is what Pulseaudio is doing a interment audio bug will be pulled out because of Pulseaudio connecting and disconnecting more often to the audio device. Please note where you stated the audio device is. Yes breaking down silicon in there is extremely bad.

    Basic fact here is ALSA does not need to expose power management controls as long as applications do the right thing. Only one problem majority don’t. Pulse-audio basically enforced correct behavior sometimes resulting in other issues showing.

    Applications being power management aware does not mean they have to directly control the power management. It means they just don’t do thing that prevent the power management system from being able to operate correctly. Pulseaudio is a power management aware application. You can save quite a large percentage of power by using power management aware. Yes while applications are connected with active buffers even if they are no sound this prevent audio devices from powering down because they have to remain prepared for incoming audio stream.

    Audio device power management logic is quite simple. If it has no streams open it can shutdown turning off amps and other things. Yes it take less than millsecond for audio device to bring its amps and other things back on-line. Audio devices sleep that often that their power management is not normally managed by the OS kernel but are hardware features. Yes the audio device hardware power management only works if the OS gives it the correct messages and the OS can only give the audio device hardware the correct state messages if the sound applications give it the correct state messages.

    Problem here is power management aware audio applications cause events to happen.

  34. ram wrote, “Do look into NetJACK”.

    It looks promising. Easy to install with Debian. No immediate joy, but no crashes or error messages either. I will get familiar with jack and then add the networking. There are several recipes on the web. Now to rid my LAN of pulseaudio.

  35. ram says:

    Personally, and this just my personal, but professional, opinion: I would kill off Pulse Audio and use JACK to transfer the audio to clients. That is what we do in my shops.

    Do look into NetJACK.

  36. oiaohm wrote, ” The fact that Pulseaudio is power management aware means running it a set of hardware errors will appear that will not be appearing running anything else Audio.”

    pulseaudio runs sound through ALSA which is not aware of power at all. That’s all controlled by the kernel which knows nothing of pulseaudio. That’s at a different layer. There’s no fan or temperature sensor on AC97 audio chips they are part of VT8235 “South Bridge” and the kernel is not going to allow pulseaudio to mess with that.

  37. oiaohm says:

    Robert Pogson
    The odds are very high against a chip failing just when I’m firing up ALSA and pulseaudio.
    pulseaudio being power management of ALSA device aware and items like esound and xrdp and Nx cleint are not.

    The odds are that the audio chip power management was failed all along.
    http://arunraghavan.net/2012/01/pulseaudio-vs-audioflinger-fight/
    Pulseaudio is not just about being a sound server. Pulseaudio is about being a power effective sound server on top of everything else. The fact that Pulseaudio is power management aware means running it a set of hardware errors will appear that will not be appearing running anything else Audio.

    Its even money that a audio hardware error will turn up as soon as you fire up Pulseaudio. In fact with Pulseaudio in jessie its more likely to be hardware fault.

    Power management items being dead are the most common causes of old thin-clients jamming up.

    Power effectiveness is not without its pain.

    one virtual machine per real machine on my network
    No you don’t need one per real machine on the network. You thin client image can be configured to run in qemu and on the real client. When you have a thin client failure might be hardware it pays to put that thin client image in qemu and see if it replicates the same fault. Normally for thin clients you have a master image with all thinclients using that image or minor variations from it.

    Its very easy to blame pulseaudio for items that turn out to be broken hardware.

  38. oiaohm wrote, ” If you do proper QA on your thin client images dead clients “.

    For one thin client, the image is what runs on that client. The odds are very high against a chip failing just when I’m firing up ALSA and pulseaudio. I’m a little short of RAM to have one virtual machine per real machine on my network.

  39. oiaohm says:

    Robert Pogson when I have being dealing with tftp booting thin clients. I set up qemu to run a pretend client. Reason qemu is software so if there is some distribution issue it shows up in qemu as well as real hardware clients. You can even setup to intentionally disrupt network traffic on the qemu.

    This allows you to have you basics right. You may be able to play music for hours because a network disruption has not happened.

    After returning to this matter, I found the thin client had dead audio-hardware.
    Robert Pogson until this point you did not notice you had a dead client. At this point you have already downgraded. You would have noticed a lot quicker if you had set up a qemu emulation of a client on the beast as it would most likely been working perfectly and the dead client would have been playing up.

    This is QA. If you do proper QA on your thin client images dead clients are in face because you can trust the images you are deploying to behave particular ways and if its not behaving those was you have a thin client issue.

    vlc and mplayer themselves does not trigger a device reset in case of lost connection instead fills the remaining buffer space with white noise. Pulseaudio does reset buffers followed by suspend on audio device in case of lost connection in default mode. What Pulseaudio does is to Alsa spec and results in lower power usage but it requires the audio hardware to be properly working. Using rdp or nx without pulseaudio thin client side instead of X11 with Pulseaudio would have also resulted in a VLC/mplayer style handling of the Audio. Tweaking the right pulse-audio settings also would have resulted in it not reseting the audio device and causing it to die.

    DrLoser there was a strange reset step in the older Linux kernels. Request what a audio device is then reset it before sending initializations instructions. This was never really required on real hardware. Required for some strange odd reason on very old versions of vmware workstation of course no one today should be using this. Resulted in some audio devices not starting at all but these devices had a defect. Debian Wheezy kernel is the last Debian kernel to include this strangeness.

    And this weird belief that a driver is “improved” if it gets “some sound” out of a “dying audio.”
    Really some sound out is a improvement. It is possible to direct pulseaudio to behave like vlc and mplayer and fill buffers out and not just reset them out of existence. Or just use RDP or NX on that client straight to alsa no pulseaudio. Basically there is no need to drop back to unsupported esound. Of course pulseaudio still runs server side when using RDP or NX.

    In 2014, however, it’s just a completely pointless waste of time.
    Not exactly pointless. A client who audio has developed lack of means to power manage/reset it self properly with pulseaudio. Pulseaudio can be configured not to attempt to power manage it or just use NX or RDP client on it without pulse-audio and it will operate almost perfectly for at least 3 years. The Almost Perfectly is so close to perfect that the users does not know anything wrong. Power bill is just a little up on what it should be. Client will be running about 1 to 2 percent more power consume. So careful accountants/ceos watching power-bills might ask a few questions when you get to 20 to 30 with this problem.

    DrLoser thin clients have annoying operational life of 10 years perfectly to about 15-20 years with some defects. Most of those some defects are so small user will not notice the administrator just has to configure a few things. NX and RDP clients using direct alsa on thin-client are better on older question mark status thin clients as long as you don’t care about power bill.

    Remember you get volume discounts ordering thin clients. It is cheaper to order 20 than its is to order 10. So being able to delay replacement has financial reasons.

  40. I doubt it’s the hardware. I just played Miley Cyrus for hours with no crash of vlc streaming from Beast to mplayer on the thin client. No pops or chirps. Just good sound.

  41. DrLoser says:

    Reality is annoying Jessie is likely due to improved drivers to get some sound out of a dieing audio where a Wheezy kernel gives up completely due to a unrequited reset step removed in newer kernels.

    And there we have it in a nutshell. Diagnosis (by oiaohm) from afar. The convenient fall-back that “I have a horible felling another hardware failure got you,” which is probably as good a guess as any considering the patient’s known tendency to “repurpose” hardware that no longer works.

    And this weird belief that a driver is “improved” if it gets “some sound” out of a “dying audio.” Really?/b> If I were stuck on Palau as a US navy wireless operator whose ship had just been sunk in, say, July 1944, this would be a terrifically important feature. Maybe even life-saving.

    In 2014, however, it’s just a completely pointless waste of time.

    I like the “unrequited reset step,” however. It brings a touch of Mills and Boone into the sad, shabby world of a complete state machine failure.

  42. oiaohm says:

    I have a horible felling another hardware failure got you.

    But, it seemed that after any interruption, mplayer was denied further connection. It went from perfection to “connection denied”.

    That behavior from jessie using pulseaudio based solutions trace to something super annoying unstable/dieing audio chips that will not reset. After a dropped connection and connection restarted pulse-audio will clear the buffers if sound card cannot reset it will die.

    Reality is annoying Jessie is likely due to improved drivers to get some sound out of a dieing audio where a Wheezy kernel gives up completely due to a unrequited reset step removed in newer kernels.

    The version of pulseaudio in Wheezy is kinda stuffed.

    Some of the exploding ram explosion is using modern X11 over network sending stacks of images. Remember fonts are these days rendered client side not server and sent across X11 protocol as images. RDP or NX has some major advantages.

    The main reason why I not worried about wayland not being a network protocol is the nasty reality that X11 is no longer suitable for light ram clients. Result is NX or RDP are the two that work. Yes using RDP or NX will see the client ram consume drop massively as well as network traffic reductions.

Leave a Reply