Message boards : LHC@home Science : Sorry CERN and LHC, but it needs to be said...
Message board moderation

To post messages, you must log in.

AuthorMessage
Shawn H. Hall

Send message
Joined: 18 Oct 16
Posts: 6
Credit: 231,270
RAC: 0
Message 33096 - Posted: 20 Nov 2017, 19:13:30 UTC
Last modified: 20 Nov 2017, 19:16:25 UTC

A user named "Yeti," I believe, took the time to crank out a super helpful checklist of what to do WHEN your LHC project craps out. It is in the Atlas messages area. And it is helpful. Unfortunately, it also points out a very important issue that I have been dealing with myself for the last few MONTHS.

Yeti is an animal...literally. I think. Isn't a Yeti an "Abominable Snowman?" Uh-huh! Well, in any case, that checklist was a bear...uh, "Yeti" to write. And about half of the information on the page didn't work...for me, at least. Plus, the CPU tester that it references says that I have an incompatible OS and system?!?!?! I have an Intel i3 virtual quad-core CPU with an Intel chipset and an Intel CPU. Also, the OS is new, with an up-to-date copy of Windows 10 on it. Consequently, and I have no idea what that darn tester was saying, my computer IS compatible, and it has far more resources than are needed to do a few of ANYTHING.

Moreover, my LHC DID work at one time, but something changed and I don't know what, but everything went wrong all at once, and *poof!* no more LHC...ANY LHC projects. It simply, and suddenly stopped working a few months ago. This set me on an odyssey of uninstalls/re-installs, unlinking/relinking/resetting, tweaking and manhandling both BOINC and my system, upgrading/downgrading/status quo installing and more. I even scoured the message boards and forums here and for my other projects to see if anything was incompatible. I found MANY similar issues, but nothing helped until user "maeax" pointed me to Yeti's checklist after a couple of months of fruitless searching and trying anything and everything that I could think to try.

So, here we are. Just Yeti and me working hard. Then, after checking off the parts of the Yeti checklist that I could, I instantly got Sixtrack working again after all this time. I was/am hoping for more, but with a majority of that graduate-level checklist left to run through, and re-check, and re-test, and re-edit, and re-start, and tweak, and then start all over again, so if that is all that I get, I will quit and call it a victory.

Thanks for the information, Yeti, but I have to say that in the past, BOINC used to do all of this work for us, while we donated our CPU cycles and some minor "tweak" time, plus a LITTLE, necessary monitoring time to the cause. What Yeti's checklist more or less now proves, without saying so, is that handling the LHC WUs is nearly a full-time job--ESPECIALLY if one wants ALL of the project apps to run, with as much data throughput as is possible, and with as many WUs cycling through as one's CPU is capable of handling, crunching through them as fast as possible...as long as one monitors it relentlessly, tweaks it relentlessly, measures the results frequently, and basically gives over a good portion of one's life to LHC.

And here's the thing: that is a huge shame. This turn of events is going to chase all but the most computer-literate nerds off of the project. The whole point of BOINC was to allow computer-ILLITERATE citizens to donate their CPU cycles over to the program, with little-to-no computer skills necessary, and with little-to-no intervention necessary, all for a noble cause. I fear that now the "baby will be tossed out the window with the bathwater" when the average Joe or Josephine finds that they have neither the time, the inclination, nor the expertise to get or to keep LHC running. And that does not bode well for the project as a whole.

Truly, and as much as it is a point of pride for me to participate, I will have to drop LHC if it consumes any more of my time. Consequently, while I know that the folks at CERN are already busy solving the riddles of the universe, whoever coded this mess needs to get back to a keyboard--with a few dozen of his/her closest friends. Then, once this program is user-friendly once again, people will use it more and more, and flock to it like the old days. And wouldn't that be nice, Yeti? Sure it would.
ID: 33096 · Report as offensive     Reply Quote
Profile Magic Quantum Mechanic
Avatar

Send message
Joined: 24 Oct 04
Posts: 1114
Credit: 49,501,728
RAC: 4,157
Message 33097 - Posted: 21 Nov 2017, 3:58:12 UTC

Well for over a decade this was the LHC SixTrack project and for many years members like Yeti and myself ran thousands of alpha and beta tests of the VB tasks to get them to work for the public.

I have been doing that daily for the last 7 years

There are just the basics for running the VB tasks and once you get them running it is easy and in all my years the only problem was the Cern server demanded me to have a fast enough ISP speed (which I deal with daily)

True that the current Sixtracks are fast and easy and the public usually has no problems with those other than when we quickly run out of WU's (something we have had to deal with since 2004)

I have never had a chance to see any of your VB tasks stderr's so I don't have any idea what your problem was........BUT you are welcome to just run SixTracks when we have them as long as you do run them.

This is my 14th year here 24/7 (and Yeti and I spend lots of time and money doing this)
Volunteer Mad Scientist For Life
ID: 33097 · Report as offensive     Reply Quote
tullio

Send message
Joined: 19 Feb 08
Posts: 708
Credit: 4,336,250
RAC: 0
Message 33099 - Posted: 21 Nov 2017, 7:29:44 UTC

I was an Alpha tester at Test4Theory@home on my Linux box. Now, after the so called "consolidation", the only LHC@home projects running on it, another Linux box and a Windows 10 PC, are Atlas@home and SixTrack. All others fail. I can ping Condor but it does not run any job and quits. No reason has been found for this behavior. VirtualBox is 5.2.0 on all 3 PCs.
Tullio
ID: 33099 · Report as offensive     Reply Quote
Profile Nils Høimyr
Volunteer moderator
Project administrator
Project developer
Project tester

Send message
Joined: 15 Jul 05
Posts: 242
Credit: 5,800,306
RAC: 0
Message 33101 - Posted: 21 Nov 2017, 8:10:30 UTC - in response to Message 33099.  

It is true that running applications under VirtualBox with BOINC is more complicated than standard BOINC applications like Sixtrack.

Thanks to all the help from volunteer testers and contributors out there like you, the process now is much smoother, but there can still be issues that we cannot automate, due to the way the computer OS is laid out.

Regarding the firewall settings, please check our FAQ on firewall ports to run VM applications.
ID: 33101 · Report as offensive     Reply Quote
Profile Laurence
Project administrator
Project developer

Send message
Joined: 20 Jun 14
Posts: 372
Credit: 238,712
RAC: 0
Message 33102 - Posted: 21 Nov 2017, 13:00:03 UTC - in response to Message 33099.  

I was an Alpha tester at Test4Theory@home on my Linux box. Now, after the so called "consolidation", the only LHC@home projects running on it, another Linux box and a Windows 10 PC, are Atlas@home and SixTrack. All others fail. I can ping Condor but it does not run any job and quits. No reason has been found for this behavior. VirtualBox is 5.2.0 on all 3 PCs.
Tullio


Can you connect to vccondorce02.cern.ch on port 9618 from your Windows machine? If you open the following link in your browser you should get a connection reset error

http://vccondorce02.cern.ch:9618
ID: 33102 · Report as offensive     Reply Quote
Profile Laurence
Project administrator
Project developer

Send message
Joined: 20 Jun 14
Posts: 372
Credit: 238,712
RAC: 0
Message 33103 - Posted: 21 Nov 2017, 13:10:14 UTC - in response to Message 33096.  

It would be great if this was easier but unfortunately we need to use virtualisation. The existence of Yeti's checklist does mean there is a fundamental issue with usability and we are trying to address this together with the BOINC community. See this presentation from the recent BOINC workshop.
ID: 33103 · Report as offensive     Reply Quote
Jim1348

Send message
Joined: 15 Nov 14
Posts: 602
Credit: 24,371,321
RAC: 0
Message 33104 - Posted: 21 Nov 2017, 13:13:53 UTC - in response to Message 33096.  

I have had the opposite experience. LHC overall, including the VirtualBox projects, has gotten much more reliable over the last year or two. I run mainly Linux now, and they are dedicated machines, but once you get them set up, they just work for me. I expect you are having a network problem, as Laurence suggests.
ID: 33104 · Report as offensive     Reply Quote
tullio

Send message
Joined: 19 Feb 08
Posts: 708
Credit: 4,336,250
RAC: 0
Message 33106 - Posted: 21 Nov 2017, 14:49:05 UTC - in response to Message 33102.  
Last modified: 21 Nov 2017, 14:56:34 UTC

I was an Alpha tester at Test4Theory@home on my Linux box. Now, after the so called "consolidation", the only LHC@home projects running on it, another Linux box and a Windows 10 PC, are Atlas@home and SixTrack. All others fail. I can ping Condor but it does not run any job and quits. No reason has been found for this behavior. VirtualBox is 5.2.0 on all 3 PCs.
Tullio


Can you connect to vccondorce02.cern.ch on port 9618 from your Windows machine? If you open the following link in your browser you should get a connection reset error

http://vccondorce02.cern.ch:9618

This is what I get on the Linux box
Same on the Windows 10 PC
THE CONNECTION WAS RESET
The connection to the server was reset while the page was loading.

The site could be temporarily unavailable or too busy. Try again in a few moments.
If you are unable to load any pages, check your computer’s network connection.
If your computer or network is protected by a firewall or proxy, make sure that Firefox is permitted to access the Web
ID: 33106 · Report as offensive     Reply Quote
Profile Magic Quantum Mechanic
Avatar

Send message
Joined: 24 Oct 04
Posts: 1114
Credit: 49,501,728
RAC: 4,157
Message 33113 - Posted: 22 Nov 2017, 2:05:11 UTC

I posted this on another thread but here it is again for any of you using Linux and the newest version of VB and Extension Pack to check.

https://www.virtualbox.org/wiki/Downloads

Update Nov 3 2017: The Guest Additions image with the 5.2.0 release had problems with a number of Linux guest systems. Please try this image which we believe fixes several of these.
Volunteer Mad Scientist For Life
ID: 33113 · Report as offensive     Reply Quote
tullio

Send message
Joined: 19 Feb 08
Posts: 708
Credit: 4,336,250
RAC: 0
Message 33117 - Posted: 22 Nov 2017, 13:18:31 UTC - in response to Message 33113.  

Atlas tasks find no problem with VB 5.2.0 on any of my PCs, Linux or Windows.
Tullio
ID: 33117 · Report as offensive     Reply Quote

Message boards : LHC@home Science : Sorry CERN and LHC, but it needs to be said...


©2024 CERN