Download problems

[Crazybob]

Is anyone else having problems with some of the files for FH not downloading with HTTP errors? I keep getting these and they sit around for a couple of hours and then get aborted.

[STE\/E]

Is anyone else having problems with some of the files for FH not downloading with HTTP errors? I keep getting these and they sit around for a couple of hours and then get aborted.
Yup all the time ... One of the worst run Projects out there ...

[Fogle]

What Ste\/e said +1 :mad:

[Mike029]

What Ste\/e said +1 :mad:

+1

[bmwr606]

+1

[denim]

I am constantly aborting and updating the project.

[coronicus]

I am constantly aborting and updating the project.

Same drives me nuts. Actual i kill freehal during chalenges cause it tends to hijack the cpu from time to time. Also always having to cancel task on reboot cause it hangs up when boinc closes. Just a little rant :ar!

[Fire$torm]

FH has been doing well on my systems for the past two weeks or so. And though this project can frustrate me at times I try not to be critical since it is being run by one individual teenager who just started his first year at university.

[Slicker]

FH has been doing well on my systems for the past two weeks or so. And though this project can frustrate me at times I try not to be critical since it is being run by one individual teenager who just started his first year at university.

I think of it this way.... We gave a junior congressman whose prior job was arranging social gatherings a crack at running the entire country and being commander in chief. By comparison, letting a college freshman run a BOINC project seems trivial. So, I agree we should cut him some slack. ;););)

[zombie67]

I look at it like this: It costs me nothing to run it. Not in CPU, not in me having to do anything. If the tasks don't d/l, or error out, no loss. If the tasks work, then I get "free" credits. Only upside. This team (and I) have made 13 million free credits that way, off of my machines alone.

[denim]

We would have all had mega credits by now if it was still running the way it used to with multiple WU's at a time.

[zombie67]

We would have all had mega credits by now if it was still running the way it used to with multiple WU's at a time.

I'm not sure I follow. I am running 12 non-cpu-intensive tasks at a time on my 12 core machine (for example).

[STE\/E]

I'm not sure I follow. I am running 12 non-cpu-intensive tasks at a time on my 12 core machine (for example).

He's talking about when multiple VM's where allowed, Guys were getting 100k+ a day Per Box from the NCI Wu's & went Ballistic when Tobias restricted the VM's to 1 Per Box. I still think some of the Guys fudge on the VM's at the Project, when I was on the SICI Team 1 Guy couldn't come near me at a regular CPU Project but would put out 5 Times what I would at the FreeHAL Project ... go figure

[artemis8]

I've been having a lot of issues with FreeHal recently where it tries to upload and fails and then waits and tries again, and does it for several hours and since it never downloads more until after it has uploaded and reported stuff my computer doesn't run anything for Hal in the meantime. Is this common?

[GalaxyIce]

I've been having a lot of issues with FreeHal recently where it tries to upload and fails and then waits and tries again, and does it for several hours and since it never downloads more until after it has uploaded and reported stuff my computer doesn't run anything for Hal in the meantime. Is this common?

Yes it's a common problem for all. I've been getting the same. The FreeHal server goes down or there is a problem there which can last for days, weeks, once for 6 months. Well, they took the system down for some months one time. Just have to wait until it will take updates again.

[GalaxyIce]

FreeHal is up and running again and accepting updates again.

[Crazybob]

Thanks for the update. Now to get to all the boxes.:)

[kaptainkarl1]

I'm not sure I follow. I am running 12 non-cpu-intensive tasks at a time on my 12 core machine (for example).

I would love to be running more than 8 at a time on my multi core machine. I have it set to run 1 instance per core. Why do I only receive/run 8 work units at a time on a 24 core system?

[Fire$torm]

I would love to be running more than 8 at a time on my multi core machine. I have it set to run 1 instance per core. Why do I only receive/run 8 work units at a time on a 24 core system?

By any chance are you using the cc_config.xml file? If so do you have a line in there ---> <ncpu>8</ncpu> ?

[kaptainkarl1]

By any chance are you using the cc_config.xml file? If so do you have a line in there ---> <ncpu>8</ncpu> ?

I will check it out as soon as I get home. Thanks F$

Edit Nope not using that File F$ any other ideas?

[Fire$torm]

I will check it out as soon as I get home. Thanks F$

Edit Nope not using that File F$ any other ideas?

The only thing left to try is to reset the project and then restart BM. Give it try and take a look at the messages sent by FreeHal. Is it reporting any errors?

[Crazybob]

I know that this is a young kid running this thing, but it is very frustrating trying to keep it going. I noticed that on this 8 core, sometimes 3 or more wu's will not DL properly. The other 5 run while it tries to DL the other files. I found that if I keep hammering the server with retries, I can get it to DL. That works fine for machines with 1 or 2 cores, but for multi-core machines, sometimes a few WU's start to run while the others are hung up and if you get the files to DL, the machine will wait until all WU's are complete before asking for more work. I have been aborting the ones that get hung up so that it will ask for the full amount of work when updating. It probably has something to do with the LAN Nazis around here, but haven't had the time to look deeper into it.

[spingadus]

I know that this is a young kid running this thing, but it is very frustrating trying to keep it going. I noticed that on this 8 core, sometimes 3 or more wu's will not DL properly. The other 5 run while it tries to DL the other files. I found that if I keep hammering the server with retries, I can get it to DL. That works fine for machines with 1 or 2 cores, but for multi-core machines, sometimes a few WU's start to run while the others are hung up and if you get the files to DL, the machine will wait until all WU's are complete before asking for more work. I have been aborting the ones that get hung up so that it will ask for the full amount of work when updating. It probably has something to do with the LAN Nazis around here, but haven't had the time to look deeper into it.

I would just not worry about it and let it do its thing. It will correct itself.

[kaptainkarl1]

The only thing left to try is to reset the project and then restart BM. Give it try and take a look at the messages sent by FreeHal. Is it reporting any errors?

Weird. Reset and restarted. Ended up with 13. Should be getting 23. Thanks for the input. I will rinse and repeat tomorrow if I don't receive more overnight.

[kaptainkarl1]

Went back to getting 8 again. What the ...? Anyone have any other ideas? Hate losing all them free stones.

[Duke of Buckingham]

http://4.bp.blogspot.com/-guOv6Jqjlio/TlhQU0srEiI/AAAAAAAAAxM/qTkKLBB5KkI/s640/lolcat_by_mooglecakezz-d366e9i.jpg

[Fire$torm]

Hmmmm... This is a wild guess but maybe its due to memory limitations? If you have a ton of RAM then check your Disk and Memory Settings in BOINC Manager under Preferences.

[kaptainkarl1]

What is the preferences overide file and how do I make it stop?

[Fire$torm]

What is the preferences overide file and how do I make it stop?

It's the locally stored file of the preferences you set in BOINC Manager. What it overrides are the settings from the web based BOINC Account Manager aka BAM. So you really DON'T want it to stop unless you only use BAM.

Here are my desktop settings for Disk & Memory Usage:
Disk Usage

Use at most 100.00 Gigabytes disk space

Leave at least 5.00 Gigabytes disk space free

Use at most 75.00 % of total disk space

Tasks checkpoint to disk at most every 60 seconds

Use at most 85.00 % of page file (swap space)


Memory Usage

Use at most 80 % when computer is in use

Use at most 90 % when computer is idle

[Check mark] Leave applications in memory while suspended

[kaptainkarl1]

Thanks FS I figured it out. Anytime you suspend a project in the manager it defaults to those settings. Still only getting 8 units and my settings were very close to yours.

[kaptainkarl1]

Strange if I reset the project and shut down boinc it will load 13 wu's then it goes back to getting 8. Maybe I will check w Rgathright on this he has the same rig as I do.

[Fire$torm]

Strange if I reset the project and shut down boinc it will load 13 wu's then it goes back to getting 8. Maybe I will check w Rgathright on this he has the same rig as I do.

Duh.... good idea!

[kaptainkarl1]

Thanks again F$ for trying to teach this old dog a thing or two.

[c303a]

I have not been able to get any work from Freehal for awhile now. I have reset the project and I have detached from it and still cannot get any work. Any ideas.

[Fire$torm]

I have not been able to get any work from Freehal for awhile now. I have reset the project and I have detached from it and still cannot get any work. Any ideas.

hmmmm. All my boxes have work. Maybe check your firewall settings? If your using something not from M$ then make sure Windows Firewall is still disabled. I have read of instances were Windows gives the user an obscure popup security warning and then the user just clicks ok. That of course turns on Windows Firewall which will cause problems.

[coronicus]

Just turned freehal back on and was able to get work for all the computer so something is up..

[c303a]

3/16/2012 10:42:58 AM | FreeHAL@home | update requested by user
3/16/2012 10:43:01 AM | FreeHAL@home | Sending scheduler request: Requested by user.
3/16/2012 10:43:01 AM | FreeHAL@home | Requesting new tasks for CPU
3/16/2012 10:43:04 AM | FreeHAL@home | [error] No start tag in scheduler reply

This is all I'm getting.

[denim]

I am able to get WU's, but it seems to get hung up from time to time.

[artemis8]

This morning both my computers aren't able to send the completed wu and get a new one.

[Crazybob]

I don't think anyone's minding the store over there. Took a while this morning to get any work sent. Kept getting the Christmas message. Also, I think it's the firewall here at work giving me a hard time. On 1 machine this morning, I had to have it update 35 times in a row to get a file to DL. Keep getting HTTP error, but eventually it will go through. Couple things to remember..... Freehal changed their URL before, so reseting wouldn't fix that, detaching would. Freehal wait's untill all you work is finished before asking for new work. So if 6 finish and 2 have an hour left, it will wait the hour.