Hardly fast though. A 80MB file took about 1.5 hours, and I had a few of them. But, I don't know if the bottleneck was the Telecentre's Internet connection or Ibiblio -- maybe the latter.
Part way through, I tried to run a second instance of gFTP, to upload the kernel source to puptrix.org. However gFTP revealed yet another bug, that a second instance does not work properly.
So, I resorted to logging into the online cpanel file manager at Hostgator (host for puppylinux.com) -- but not after considerable trouble as they had changed my password (they had changed everyones password that they considered to be too weak). Anyway, I used that to upload the kernel source to puppylinux.com, and while that was happening the ibiblio upload became even slower, which indicates the bandwidth limitation is at the Telecentre end.
I was there for 5 hours! And still had one file to go, which I decided to upload from home. Well, it's uploading right now, a 91MB file, estimated time about 5-6 hours total, and 3 to go -- as long as the satellite stays in-sync that long, which is doubtful.
4.1alpha iso builds
Enough whinging. I have uploaded 'puppy-4.1alpha1-seamonkey-cutdownmodules.iso' and 'devx_401.sfs', that you can find here:
The file 'puppy-4.1alpha1-seamonkey.iso' is at the time of writing still uploading. Until you see it as 95,940,608 bytes (91.5MB), it isn't fully there yet.
Puppy is a "cutdown" distro, and I have always had reservations about including all the modules. The "cutdownmodules" build has a selection that is likely to satisfy the vast majority of users, though it may need a bit of tweaking to get right. I am interested in working on this to get it as the main official Puppy build.
I have compiled the 126.96.36.199 kernel with many more modules enabled, adding about 2MB compressed, compared with the 188.8.131.52 kernel. That is the main reason the full build is 91.5MB.
Regarding the kernel source, I've also got all the third-party modules, plus some extra doc files, and intend to upload all of that to puptrix.org. For now, the patched kernel source and the .config file can be found here:
Regarding the event management in Puppy, I have started to write a web page on that, will complete and upload that tomorrow probably. For now, you can examine what has been going on in /etc/rc.d/rc.modules, rc.modules2, and /sbin/pup_eventd.
There's a lot that is preliminary -- for example, some code in pup_eventd works with JWM but it should really be generalised to work with other trays like lxpanel. Ditto for Pmount.
Note that the 'init' script has not yet been modified to make use of the fact that all the modules are present at that early stage of bootup. I still have to think about some issues. The only extra code in the init script (compared with pup 4.00) is for handling moving of the modules to the main unionfs filesystem.
Comments:Posted on 26 May 2008, 18:47 by GratnessGuru
Zero length "md5sum" file.
26-May-2008 03:16 0 < Zarro bytes.
Posted on 26 May 2008, 19:17 by BarryK
Heh, heh, before anyone tells me, I have just discovered a bug at line 81 in rc.modules.
Posted on 26 May 2008, 19:51 by nfisher
Actually it would be nice for me if the pup_eventd didn't interact with X-windows at all, reason being that Grafpup is multi-user and anything it requests to be put into the tray, or programs it launches, will be run as root even if the user who is logged in is not root. But I have resigned myself to the fact that I may have to do a lot of re-writing to separate the code into two daemons, a backend which does things like load modules and add device nodes, possibly adding fstab entries, and a frontend which can be run with a user's session to add desktop icons when removable media is inserted, or other such functions.
Anyway, my 2c. I've just been waiting to look at the code so I can get a more detailed idea of how it works now so I can possibly mimic it in my own implementation.
Posted on 26 May 2008, 20:19 by Rob
Slow upload speed
Barry, is there anyone in your area with broadband? It would seem to me a much simpler solution to your problem would be to give someone else access and to get them to do the final upload. You could send it to them on cd. That way a small change like that could be made in an instant. I'd offer, but I think the opposite side of the planet is about as far away from you as I can be, although I do have 24mb access (1.3 upload i think) and faster than that at uni!!!
Posted on 26 May 2008, 21:34 by BarryK
4.1alpha full size
Ok, the full-size iso has uploaded. 91.5MB in 5 hours!
i was tossing up whether pup_eventd should be started earlier, from rc.local0, so it doesn't have to be killed when X quits.
My satellite connection is broadband, ditto for the Perenjori Telecentre. There's broadband and there's broadband....
Next upload I do will be from a relative's place in Perth, lovely adsl2.
Posted on 26 May 2008, 23:08 by Prit1
Are we supposed to report bugs that we see or no?
I understand that this is just a preview for developers.
Anyway, just reporting what I did see. Bootup, saving session and all went smooth for me. Abiword, Gxine and all worked. Network didnt save for next session, Closing pidgin crashed X.
Posted on 27 May 2008, 2:12 by veronicathecow
Hi Barry, no joy with this version either. It sticks on booting the kernel in frugal on my intel D201GLY2 MB. Perhaps I've been stupid with my GRUB?
kernel (hd0,7)/boot/vmlinuz BOOT_IMAGE=linux root=/dev/sda8 acpi=on resume=/dev/sda6 splash=silent vga=788
title Puppy Linux 401 frugal
kernel /puppy401/vmlinuz pmedia=satahd psubdir=puppy401 acpi=off
Posted on 27 May 2008, 2:26 by GreatnessGuru
"There's broadband and there's broadband...."
My little farming town here has DSL-512,
a vast improvement over the DSL-256
we had until three years ago!
(Much higher speeds at extra cost.)
I'm now downloading
at ~60KB/sec, ~25 minutes total.
Next year we get fiber-to-the-home.
So how your Telecenters, Barry,
can be so slow I can't imagine.
If you wish, I think most of us would be happy
if you wanted to take a couple of days
to upload things sometimes.
Spend your upload time working on
your windpower project
instead of wasting fuel and time
driving and sitting around.
Inwood IA USA
Pop: "875 with 365 housing units"
Posted on 27 May 2008, 2:44 by drongo
Barry, Busybox supports tftp. I don't know if you included it in Puppy as I'm currently on a Windows XP (work) machine and I don't know if you included everything in Busybox as you only need to include what you want. tftp uses udp NOT tcp, this means that you shouldn't get slow speeds over satellite as there is none of that ACK and window resizing nonsense. The downside is that if packets are corrupted/lost you get no indication. Your satellite link should have an FEC wrapper around the data. (Is it some form of DVB-S link?) This FEC wrapper should protect against a fair amount of corruption or loss. You can always verify the uploaded file by doing an md5 check on it I guess?
Might be worth giving tftp a shot? No point in paying for bandwidth you can't use.
If you are using Firefox as a browser you can use download acceleration by modifying the config file. Not sure if this works in Seamonkey.
Basically, if you want decent bandwidth over a satellite link you either need multiple tcp sessions or one udp transfer. If the udp doesn't work reliably you could always consider wrapping it in your own FEC and stripping this off after reception.
Posted on 27 May 2008, 5:20 by twointo1
zd1211rw firmware won't load
Dmesg shows an "error loading firmware zd1211rw - error -2 Usb wireless works in all Puppies since 2.17 including 4.0 beta 2.6.25. Tried unplugging and plugging back in but I get the same error.
Posted on 27 May 2008, 8:45 by kirk
ipw2200 firmware won't load
ipw2200: Detected Intel PRO/Wireless 2200BG Network Connection
ipw2200: ipw2200-bss.fw request_firmware failed: Reason -2
ipw2200: Unable to load firmware: -2
ipw2200: failed to register network device
During boot it takes about two minutes to get past trying to load this firmware.
Posted on 27 May 2008, 8:45 by kirk
Installed hotplug from 4.0, ipw2200 working again.
Posted on 27 May 2008, 10:16 by wildpossum
tftp is actually the wrong direction to take with a high latency link as it's a send-ack protocol. You don't get the streaming that you do with TCP's window.
Perhaps something can be done to tweak TCP over satellite, or are there no knobs available, Barry?
Posted on 27 May 2008, 16:20 by BarryK
Satellite upload speed
My satellite upload speed is limited to 64K/sec, that's dialup speed, and there's nothing I can do about that. I am getting slower than that because of technical issus, but even if I tweaked it somehow it is still only going to be approaching 64K/sec.
See this url:
I'm only paying AU 39.95, I would have to go up to 79.95 to double the speed.
Perenjori Telecentre is on a wireless connection:
...I haven't asked them which plan, but I know they have serious financial problems so I doubt that it's one of those top plans.
Posted on 27 May 2008, 21:54 by Jesse
I have been having a go with 4.01alpha1 and I run into a problem when I copy the contents of devx.sfs to the /. The computer hangs, no mouse movement.
I have 2 Gig ram, and the devx files are about 352Mb, so plenty of space. I boot pfix=ram.
I also notice that there is a .wh. file in the 401 devx.sfs file /usr/share/idl/seamonkey-1.1.8/.wh.*
I've run the testmem program, and no issues.
Is it a driver or the kernel? anyone else have this issue?
Posted on 27 May 2008, 24:33 by GreatnessGuru
Re: 184.108.40.206 unstable?
Yes. I have. Puppy 4.00.
I thought it might be the kernel,
fresh 220.127.116.11 release, you know.
Haven't tried it with Puppy 4.01.
Posted on 28 May 2008, 4:25 by GreatnessGuru
Re: Satellite upload speed
"My satellite upload speed is limited to
64K/sec, that's dialup speed, ..."
"BBSatLite 512kb/64kb 500MB $39.95"
"List of device bandwidths"
"ISDN Basic Rate Interface &
Primary Rate Interface data
64/128 kbit/s 8/16 kB/s"
With various overheads and whatnot
that's ~6 kBytes/sec under "good" conditions.
That's ~10% what I have with DSL-512 d/l.
26-May-2008 09:09 91M,
that took me ~25 minutes to d/l,
should take you, Barry, ~250 minutes,
over 4 hours, to u/l.
You said it took 5 hours.
That may be about as good as
can be expected.
Again, Barry, I think most of us would
be happy for you to take one-two days
now and then to upload your stuff,
just so you can stay put and
not have to waste time and fuel on
deprecated, wireless(!) "Telecenters",
as that one insists on calling itself.
It's no wonder they are going broke,
and soon out of business, I bet.
Posted on 28 May 2008, 9:50 by BarryK
Most small towns in Western Australia have a Telecentre. It's not privately owned, it's a community thing, reliant on some govt grants and community input. Ours also has the Post Office and Centrelink agencies. It has the equivalent of one full-time staff member, the rest are volunteers. Perenjori is very small, that's the main reason it is borderline viable.
Note, "re", not "er" as we use British spelling. Not entirely, we are gradually getting corrupted, for example I usually type "color" rather than the correct "colour". It doesn't help that I leave the US spellchecker in SeaMonkey!
Posted on 28 May 2008, 10:35 by GreatnessGuru
"WA Telecentre Network Home"
"Home > Telecentre Details
Telecentres in the Mid West Region"
"Perenjori Telecentre ..."
Working Paper Series
Paper No. 10
Themes and Issues in Telecentre Sustainability
RAUL ROMAN & ROYAL D. COLLE
"Themes and Issues in Telecentre Sustainability
Raul Roman & Royal D. Colle
Cornell University, Ithaca, New York ...
The initiation, diffusion and adoption of the telecentre idea has been an enormously eclectic process, largely devoid of systematic research and planning.
The approach has generally been one of pilot projects —
trying out models to see what works to achieve a diversity of objectives.
In some cases the approach has been simply entrepreneurial, with enterprising business people exploring new opportunities for profit-making.
Often mentioned but largely undeveloped is the training associated with telecentre management, an issue that relates to all of the issues mentioned.
While each of the issues deserves systematic analysis, this chapter concentrates on sustainability and training.
Based on data collected from Australia and South Africa to Hungary and Canada —
and from various project documents —
we describe some of the strategies being used to sustain telecentres.
We put this discussion especially in the context of developing nations because of the intense interest in the early 21 st century in incubating telecentres in places where individual connectivity to information access is most problematic."
Posted on 30 May 2008, 4:24 by GreatnessGuru
Re: Jesse, "18.104.22.168 unstable?"
27 May 2008, 21:54 by Jesse
I have now tried with Puppy 4.01.Alpha.1.
The problem I had with the mouse hanging
in X in Puppy 4.00 right after booting
with the /devx_400.sfs & the kernel .sfs
files is gone. No more mouse hangs.
I first booted Puppy 4.01 with the
/devx_401.sfs file only. No mouse hang.
Then I d/l /kernel-src_401.sfs and
rebooted with both .sfs files
as I had with Puppy 4.00.
I'm now using that Puppy 4.01
with those two files as I type this,
so no more mouse hangs for me in 4.01.
Posted on 30 May 2008, 5:08 by GreatnessGuru
Re: Jesse, "22.214.171.124 unstable?"
Re: Jesse, "126.96.36.199 unstable?"
I was using the old Puppy 4.00 kernel!
I had not updated the GRUB menu file
from "/boot/vmlinuz" to "/vmlinuz".
Now, with the correct Puppy 4.01
kernel, my mouse Does Hang
after reboot with both of those .sfs files.
Posted on 30 May 2008, 6:34 by GreatnessGuru
Re: Jesse, "188.8.131.52 unstable?"
A possible clue...
The previous trials, and tribulations,
all happened with a Full Install Boot,
Puppy 4.00, then upgraded to Puppy 4.01.
Just now, I went back to my main
Puppy 4.01 Frugal Install Boot, and,
when X came up, BootMgr asked about
those two .sfs files! What?
Those two .sfs files, devx and kernel
for 4.01, are at sda1:/...sfs,
while this Frugal Install is at
So why would BootMgr be asking about files
outside the /puppy401 directory it is
booting from during this Frugal Session?
Anyway, I Canceled BootMgr.
But, here's the thought/clue:
Could the Full Boots,
with those .sfs files present,
be Hanging the Mouse at the point
where BootMgr is suppose to be launched
upon discovering those .sfs files there?
I am clueless about how this is all
suppose to work with adding extra
.sfs files. Hanging the Mouse in
Full Boot mode must be a Bug, though,