New release of diskless folding suite

Come join the... uh... er... fold.

Moderators: just brew it!, farmpuma

Re: New release of diskless folding suite

Postposted on Thu Apr 10, 2008 6:27 pm

I got the USB stick version to work perfectly, but I wanted to use my computer while it folds.

So I got VMWare Player and downloaded the fold.iso and the fold64.vmx file. Everything fires up perfectly and the VM runs up until the very end, where it says "setting up instance 1".

Then is has an error message - "mount: mounting /dev/hda1 on /hda failed: No such device or address"

Soon after I get "Disk image is corrupt, not restoring".

What am I doing wrong? Is there any way to fix this?

~LumberJack~
LumberJack
Gerbil
 
Posts: 13
Joined: Thu Apr 10, 2008 6:14 pm

Re: New release of diskless folding suite

Postposted on Thu Apr 10, 2008 6:28 pm

That's just a vestigial error message. It's actually running fine.

Welcome to the TR forums 8)
Usacomp2k3
Gerbil God
 
Posts: 21323
Joined: Thu Apr 01, 2004 4:53 pm
Location: Orlando, FL

Re: New release of diskless folding suite

Postposted on Thu Apr 10, 2008 7:38 pm

Usacomp2k3 wrote:That's just a vestigial error message. It's actually running fine.

Welcome to the TR forums 8)


That was fast. Thanks!

~LumberJack~
LumberJack
Gerbil
 
Posts: 13
Joined: Thu Apr 10, 2008 6:14 pm

Re: New release of diskless folding suite

Postposted on Thu Apr 10, 2008 8:02 pm

LumberJack wrote:
Usacomp2k3 wrote:That's just a vestigial error message. It's actually running fine.

Welcome to the TR forums 8)


That was fast. Thanks!

~LumberJack~

What can I say. I have no life :wink:
Usacomp2k3
Gerbil God
 
Posts: 21323
Joined: Thu Apr 01, 2004 4:53 pm
Location: Orlando, FL

Re: New release of diskless folding suite

Postposted on Thu Apr 10, 2008 9:28 pm

Usacomp2k3 wrote:
LumberJack wrote:
Usacomp2k3 wrote:That's just a vestigial error message. It's actually running fine.

Welcome to the TR forums 8)


That was fast. Thanks!

~LumberJack~

What can I say. I have no life :wink:

Nah, you are just TR's posterboy PWer. :P
Image
The Model M is not for the faint of heart. You either like them or hate them.

Gerbils unite! Fold for UnitedGerbilNation, team 2630.
Flying Fox
Gerbil God
 
Posts: 24580
Joined: Mon May 24, 2004 2:19 am

Re: New release of diskless folding suite

Postposted on Thu Apr 10, 2008 9:37 pm

Flying Fox wrote:
Usacomp2k3 wrote:What can I say. I have no life :wink:

Nah, you are just TR's posterboy PWer. :P

Shh. My wife might hear :wink:
Usacomp2k3
Gerbil God
 
Posts: 21323
Joined: Thu Apr 01, 2004 4:53 pm
Location: Orlando, FL

Re: New release of diskless folding suite

Postposted on Sat Apr 12, 2008 3:37 pm

theMASS wrote:Sounds like you were running the standard client (x 2), and now notfred's software is running the SMP client as it is detecting the dual core processor. If you were running a REALLY old version this makes sense. If your "prior" version was from the last year... then yes weird =:?


Yeah, i think it was some ver. from nov.07 , before the menu was added, :o
bj0ng0
Gerbil In Training
 
Posts: 3
Joined: Thu Apr 10, 2008 4:19 pm

Re: New release of diskless folding suite

Postposted on Fri May 02, 2008 9:18 pm

notfred: On the network diskless setups. Instead of naming the configuration settings pxelinux.cfg/default, can I name the file to match the MAC address per the pxelinux standards so I can boot multiple diskless clients? I have static IPs for each host on my dhcp server. Or do I need a separate install off of /var/lib/tftpboot for each host?

Thanks
Tobit
Gerbil In Training
 
Posts: 5
Joined: Fri May 02, 2008 4:06 pm

Re: New release of diskless folding suite

Postposted on Sun May 04, 2008 9:39 pm

notfred: If you could add ... it would be nice to see a USB boot parameter that allows a static address to be assigned in the syslinux.cfg file if possible?
usa2k
Gerbil
 
Posts: 11
Joined: Fri Apr 06, 2007 12:45 am

Re: New release of diskless folding suite

Postposted on Tue May 06, 2008 7:26 pm

Tobit wrote:notfred: On the network diskless setups. Instead of naming the configuration settings pxelinux.cfg/default, can I name the file to match the MAC address per the pxelinux standards so I can boot multiple diskless clients?
Yes, it's standard pxelinux boot so you can name by IP address, MAC address or partials of either or allow it to fall through to the default.

usa2k wrote:notfred: If you could add ... it would be nice to see a USB boot parameter that allows a static address to be assigned in the syslinux.cfg file if possible?
It's actually a few parameters
  • IP Address
  • Subnet mask
  • Gateway
  • DNS server
and almost all networks support DHCP these days. It should be possible to add it though - I've raised a feature request to track it:
https://sourceforge.net/tracker/index.p ... id=1021760
notfred
Grand Gerbil Poohbah
 
Posts: 3775
Joined: Tue Aug 10, 2004 10:10 am
Location: Ottawa, Canada

Re: New release of diskless folding suite

Postposted on Wed May 07, 2008 10:46 pm

I've noticed a problem with one of my diskless network servers -- the backups are being made in the default tftpboot directory, but for some reason, the corresponding "latest" files are zero bytes. There are two clients running off of this server, and the time stamps on both "latest" files match the most recent "backup" file for each client, but the "latest" file itself is empty. The "backup" files are being created properly. On my other diskless network server, the "latest" files are working as they should (2 bytes, containing "A" or "B").

I've reinstalled the diskless code, deleted/touched/chmod'ed the "latest" files, and nothing seems to fix the problem.

Anyone else seeing this happen? Any ideas as to what's wrong?
Hyperlife
Gerbil In Training
 
Posts: 2
Joined: Thu Jan 24, 2008 10:53 pm

Re: New release of diskless folding suite

Postposted on Fri May 09, 2008 6:48 pm

Hi there :)

I just noticed that there is no syslinux install file for Vista, only for Windows NT, 2000, XP, 2003, Windows 95, 98, ME, and Linux (for the USB Stick method). I know nothing of Vista having never used it before, but i am asking on behalf of someone else. Would a file have to be created on one of the supported OS's then passed along? Or is there something else to try? (compatibility mode just came to mind.)

Thanks.
Main: Opty165 @ 289x9 | Gskill HZ 2x1gb @ 260mhz | DFI NF4 Ultra-D | x1900xt | NEO HE550 | 32bit XP Pro |
Server: E2180 @ 300x10 | 996533 @ 450mhz | GA-P35-DS3L | 7300GT | Smartpower 450 | 64bit Ubuntu 7.10 |
EllisD
Gerbil In Training
 
Posts: 2
Joined: Tue Apr 01, 2008 12:14 pm
Location: Vancouver BC

Re: New release of diskless folding suite

Postposted on Fri May 09, 2008 9:16 pm

Not sure, I've never even seen Vista and I don't have any other Windows machines - I'm a hard core Unix guy.

http://www.damnsmalllinux.org/wiki/inde ... lash_Drive suggests that on Vista you need to start cmd.exe (the command prompt) as Administrator, and then run the NT, 2000, XP, 2003 version of syslinux from my webpage.

One other alternative would be to boot in to the diskless folding via some other method (cdrom or network) with the INSTALL option enabled and then let it set up the USB stick.
notfred
Grand Gerbil Poohbah
 
Posts: 3775
Joined: Tue Aug 10, 2004 10:10 am
Location: Ottawa, Canada

Re: New release of diskless folding suite

Postposted on Sat May 17, 2008 9:36 pm

I installed using a USB stick and had ZERO problems. So I phoned my friend Chris to tell him, and he was working all afternoon to try to get it running. But he keeps getting this error.

Unpacking initramfs...<0>Kernal panic - not syncing: invalid compressed format (err=1)

He has a GA-P35-DS3L mobo. And a Q6600. He runs WinSMP perfectly in XP.

I'm wondering what is causing that. We tried USB-HDD, USB-ZIP, USB-Floppy, etc in bios. It actually runs and detects "fold64" and seems to be going along perfectly. Then *wham* "Unpacking initramfs...<0>Kernel panic - not syncing: invalid compressed format (err=1)"

Any help would be much appreciated. Thanks

Bruce
LumberJack
Gerbil
 
Posts: 13
Joined: Thu Apr 10, 2008 6:14 pm

Re: New release of diskless folding suite

Postposted on Sun May 18, 2008 7:24 am

notfred - First off many thanks for all the hard work you do on this ...

A Few of us over at [H]ardforums were talking about something the other day .. and was wondering if this is possible? I know it sounds crazy though..

Instead of booting from a USB or CD, is there a way to run this off a hard drive. Now, we know it kinda defeats the purpose of the diskless thing .. but many over there love the linux client but know 0 about linux and don't wanna run a gui linux distro.... your setup saves so much time, get the shares correct, and network settings all at one time ..

Is this even possible? or are we just crazy ...
EvilAlchemist
Gerbil
 
Posts: 28
Joined: Tue Jan 29, 2008 1:54 am

Re: New release of diskless folding suite

Postposted on Sun May 18, 2008 12:44 pm

I'm the guy mentioned back on March 30 in this thread with all sorts of trouble with the clock running wild in VMWare. Trying to run notfred's folding in Linux under VMWare on a host Windows XP SP2 machine. AMD X2 4200. Clock on the VMWare machine running crazy fast i.e. just tried it now and it had ticked off 37 minutes when in fact only 7 minutes of real time had passed. This, apart from just not being perfect, is causing actual problems in that folding deadlines pass (or appear that way to the folding software). I'm just getting back to the problem now, real life snuck in with a problem, but I do want to get this fixed so I can run notfred's, and the real frustrating part is that the identical setup is running fine on the machine next to it, an AMD X2 4000. Contrary to the conclusion reached back at the end of March here, I had already had the AMD Dual Core Optimizer installed and in fact when I tried to install it just now it told me so. Also just now I've removed and then re-installed VMWare server and re-d/led fold.iso and fold64.vmx.

The identical problem persists, any ideas, the AMD Dual Core Optimizer wasn't the fix. Frustrating, I want to fold more efficiently. Thanks.
ArVee
Gerbil
 
Posts: 21
Joined: Sun May 18, 2008 12:21 pm

Re: New release of diskless folding suite

Postposted on Mon May 19, 2008 8:06 am

LumberJack wrote: We tried USB-HDD, USB-ZIP, USB-Floppy, etc in bios. It actually runs and detects "fold64" and seems to be going along perfectly. Then *wham* "Unpacking initramfs...<0>Kernel panic - not syncing: invalid compressed format (err=1)"

That's pretty bizarre. I assume you tried copying over the files multiple times to make sure it is not just a bad copy of the initrd? Another thing to try is to boot off a folding CD with the INSTALL on and that will try to install to the first two USB sticks that it finds.

EvilAlchemist wrote:Instead of booting from a USB or CD, is there a way to run this off a hard drive. Now, we know it kinda defeats the purpose of the diskless thing .. but many over there love the linux client but know 0 about linux and don't wanna run a gui linux distro.... your setup saves so much time, get the shares correct, and network settings all at one time ..

Is this even possible? or are we just crazy ...

It's possible (although I'd need to add support for the hard drive controllers in). I'm just a little wary of doing this as I don't want it to go and ruin an OS install if someone boots off a CD or USB or similar. I've added it as an enhancement request 1967063

ArVee wrote:I'm the guy mentioned back on March 30 in this thread with all sorts of trouble with the clock running wild in VMWare. Trying to run notfred's folding in Linux under VMWare on a host Windows XP SP2 machine. AMD X2 4200. Clock on the VMWare machine running crazy fast i.e. just tried it now and it had ticked off 37 minutes when in fact only 7 minutes of real time had passed.

That sounds like more of a VMWare problem than an issue with my stuff. Have you tried Googling for "vmware clock issues"? The first hit is a PDF from VMWare about how clock syncing is done - you want to enable that as my stuff detects VMWare and stops running NTP. One other thought, check your motherboard for bad caps. I had a motherboard that worked perfectly for ages, then the time started going wonky then about 3 months later it died with bad caps. I would have thought that an X2 capable motherboard was passed the date when bad caps were around though.
notfred
Grand Gerbil Poohbah
 
Posts: 3775
Joined: Tue Aug 10, 2004 10:10 am
Location: Ottawa, Canada

Re: New release of diskless folding suite

Postposted on Tue May 20, 2008 3:01 pm

LumberJack wrote: We tried USB-HDD, USB-ZIP, USB-Floppy, etc in bios. It actually runs and detects "fold64" and seems to be going along perfectly. Then *wham* "Unpacking initramfs...<0>Kernel panic - not syncing: invalid compressed format (err=1)"


notfred wrote:That's pretty bizarre. I assume you tried copying over the files multiple times to make sure it is not just a bad copy of the initrd? Another thing to try is to boot off a folding CD with the INSTALL on and that will try to install to the first two USB sticks that it finds.


Thanks for the reply. And thanks for all you do notfred.
I hate to even bother you with dumb qustions.

notfred wrote:Another thing to try is to boot off a folding CD with the INSTALL on and that will try to install to the first two USB sticks that it finds.

Since he is using a quad. (Q6600) that will run 2 instances. Does he need to have 2 sticks plugged in? Or maybe you were just saying it would install to 2 sticks. But you only need ONE. right? Sorry for the *cornfusion* ;)
Last edited by LumberJack on Wed May 21, 2008 3:36 am, edited 1 time in total.
LumberJack
Gerbil
 
Posts: 13
Joined: Thu Apr 10, 2008 6:14 pm

Re: New release of diskless folding suite

Postposted on Tue May 20, 2008 3:18 pm

LumberJack wrote:
LumberJack wrote: We tried USB-HDD, USB-ZIP, USB-Floppy, etc in bios. It actually runs and detects "fold64" and seems to be going along perfectly. Then *wham* "Unpacking initramfs...<0>Kernel panic - not syncing: invalid compressed format (err=1)"[/quote
notfred wrote:That's pretty bizarre. I assume you tried copying over the files multiple times to make sure it is not just a bad copy of the initrd? Another thing to try is to boot off a folding CD with the INSTALL on and that will try to install to the first two USB sticks that it finds.]


Thanks for the reply. And thanks for all you do "notfred".
I hate to even bother you with dumb qustions.

notfred: "Another thing to try is to boot off a folding CD with the INSTALL on and that will try to install to the first two USB sticks that it finds."

Since he is using a quad. (Q6600) that will run 2 instances. Does he need to have 2 sticks plugged in? Or maybe you were just saying it would install to 2 sticks. But you only need ONE. right? Sorry for the *cornfusion* ;)


Only 1 stick is needed
Image
theMASS
Gerbil First Class
 
Posts: 132
Joined: Thu Sep 27, 2007 3:24 am

Re: New release of diskless folding suite

Postposted on Tue May 20, 2008 3:31 pm

Thanks TheMASS!
LumberJack
Gerbil
 
Posts: 13
Joined: Thu Apr 10, 2008 6:14 pm

Re: New release of diskless folding suite

Postposted on Tue May 20, 2008 7:26 pm

Thank you notfred, still no luck here, but appreciate the help and I've learned more than I thought I'd ever know about clock interrupts. :P I agree that the snag is very likely with VMWare.
ArVee
Gerbil
 
Posts: 21
Joined: Sun May 18, 2008 12:21 pm

Re: New release of diskless folding suite

Postposted on Wed May 21, 2008 3:03 pm

theMASS ,

Did you ever conclude your tests on the /2 vs /4 config.

I have read threw that area of the thread .. and it seems like you were getting around 10%+ boost at a min ...

I have noticed that on the 3065 ... there is not much boost from going 2x SMP ... but all others get a good boost.

I have been on WIN for a while, and just move 5 systems back to diskless...

I had forgotten how many more WU seem to be available for linux that are not that common in windows.
EvilAlchemist
Gerbil
 
Posts: 28
Joined: Tue Jan 29, 2008 1:54 am

Re: New release of diskless folding suite

Postposted on Thu May 22, 2008 1:55 am

notfred wrote:Another thing to try is to boot off a folding CD with the INSTALL on and that will try to install to the first two USB sticks that it finds.


That worked! Thank you Mr notfred. :)

For some reason my buddy's Gigabyte mobo, and mine also, didn't seem to like booting, then running from a USB stick. It booted fine, but wouldn't actually run the Linux. On my Athlon64 it boots and runs fine with just a USB stick.
So we burned a CD. Set we the bios order to CD-Rom, then USB-HDD. It booted from the CD, and saved the files to the USB flash drive like it should. This might be an all around solution to those that have trouble booting and running exclusively from a USB stick.

Thanks for all your hard work in making this folding suite notfred! :P
LumberJack
Gerbil
 
Posts: 13
Joined: Thu Apr 10, 2008 6:14 pm

Re: New release of diskless folding suite

Postposted on Wed May 28, 2008 6:52 pm

Is there any way to remove the -forceasm and -advmethods from the notfred folding suite?

Arguments: -local -forceasm -verbosity 9 -advmethods

If there is I would be greatfull...

And thanks for the great way to pxe boot into folding
Image
dierdre7
Gerbil In Training
 
Posts: 4
Joined: Tue May 27, 2008 10:13 pm

Re: New release of diskless folding suite

Postposted on Fri May 30, 2008 8:54 am

Not at the moment - I do have an enhancement request open to allow tweaking of all the options, but it will not be in the next release - that will contain a VMWare setup.
notfred
Grand Gerbil Poohbah
 
Posts: 3775
Joined: Tue Aug 10, 2004 10:10 am
Location: Ottawa, Canada

Re: New release of diskless folding suite

Postposted on Sun Jul 27, 2008 11:39 pm

EvilAlchemist wrote:theMASS ,

Did you ever conclude your tests on the /2 vs /4 config.

I have read threw that area of the thread .. and it seems like you were getting around 10%+ boost at a min ...

I have noticed that on the 3065 ... there is not much boost from going 2x SMP ... but all others get a good boost.

I have been on WIN for a while, and just move 5 systems back to diskless...

I had forgotten how many more WU seem to be available for linux that are not that common in windows.


Sorry for the extremely delayed response...

Even though on most WUs running 2x SMP clients on quads theoretically result in a 10 - 15% higher PPD than a standard 1x SMP setup, I found that my actual PPD was the same or better with the 1x SMP setup due to stability issues. My 1x setups can run for months without crashing while the my 2x setups usually needed some "help" every few days.

On machines that were/are running 2 VMs I get about a 30% increase over 1x SMP. The main reason for the increase with 2 VMs is the WUs assigned to dual core setups are higher PPD WUs. The VMs get almost exclusively 2605s and no 2665s.

This may all change now that the A2 core scales better than the A1 core.
Image
theMASS
Gerbil First Class
 
Posts: 132
Joined: Thu Sep 27, 2007 3:24 am

Re: New release of diskless folding suite

Postposted on Wed Aug 13, 2008 6:09 pm

Is there any way to get -forceasm disabled in the latest release (Aug.3)? It doesn't play nicely with -advmethods on the new _a2 core, which actually runs 50% faster (!) if it would only get through a WU without irretrievably crashing. Those who know better than I say it's the -forceasm or the combination.
ArVee
Gerbil
 
Posts: 21
Joined: Sun May 18, 2008 12:21 pm

Re: New release of diskless folding suite

Postposted on Wed Aug 13, 2008 6:35 pm

Interesting, do you have a pointer to where they discuss this? The reason I ask is because I've been running core a2 in general fine with -forceasm on my Linux (non-diskless machines). I get the odd hang at the end here and there, but I've only seen 2 of them crash in the middle of the WU over a couple of weeks of core a2s.
notfred
Grand Gerbil Poohbah
 
Posts: 3775
Joined: Tue Aug 10, 2004 10:10 am
Location: Ottawa, Canada

Re: New release of diskless folding suite

Postposted on Wed Aug 13, 2008 6:48 pm

I know there was more than this, but this and the link therein contained gets into some of it.

http://foldingforum.org/viewtopic.php?f ... =90#p47116

Glad it's working for you, I tried for three days before unchecking the Advanced Methods bullet and then the crashes stopped as _a1's came back. I'd far rather run _a2's successfully though.
ArVee
Gerbil
 
Posts: 21
Joined: Sun May 18, 2008 12:21 pm

Re: New release of diskless folding suite

Postposted on Wed Aug 13, 2008 6:59 pm

There this as well which writes to the FAH logfile when -advmethods and -forceasm are used in combination, although it's a "canned" message.

By using the -forceasm flag, you are overriding
safeguards in the program. If you did not intend to
do this, please restart the program without -forceasm.
If work units are not completing fully (and particularly
if your machine is overclocked), then please discontinue
use of the flag.

There was more detailed discussion of this topic, but I can't find it at the moment, I just remember salting it away.

Thanks for looking at it regardless, nf.
ArVee
Gerbil
 
Posts: 21
Joined: Sun May 18, 2008 12:21 pm

PreviousNext

Return to TR Distributed Computing Effort

Who is online

Users browsing this forum: No registered users and 1 guest