| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: OpenPandemics - COVID-19 Project
Thread: 2022-12-08 Update (OPNG workunits, storage update, and missing devices on Contributions page) |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 83
|
|
| Author |
|
|
binventive
Cruncher Joined: May 3, 2007 Post Count: 13 Status: Offline Project Badges: 
|
Well, right in time for the weekend, and after luckily not seeing any such problems for the last two weeks or so, it looks like download errors are back with a vengeance...  Ralf  Sigh, I noticed the same thing after being away on vacation for a week (prior to which the downloads had been working fine for a bit at least). Time to run Rosetta, SiDock, and TN-Grid again (although the work available for those projects tends to be sporadic it appears).
----------------------------------------
---------------------------------------- [Edit 1 times, last edit by binventive at Dec 10, 2022 1:20:57 PM] |
|||
|
||||
|
Paul Schlaffer
Senior Cruncher USA Joined: Jun 12, 2005 Post Count: 247 Status: Offline Project Badges: 
|
Yes, ARP is back, and so are the download issues. It would really make a lot of sense to keep this one on standby until the new storage system is up and running, so we can have a smooth running system. I don't run any ARP at the moment and yes, the download issues are back on both OPN and MCM. I also noticed I am getting about 50% MCM units which are retries of the second or third time. Probably two unrelated issues. Cheers He is not saying when you personally have ARP then you get download issues, he is saying that when ARP are available to download en masse, then the transient http errors start showing up. Magically they have just now done so after weeks of being fine for everyone. Magically the second ARP units get turned back on! I really second the proposal, the project should not give out ARP units until at least the SSD bank comes online. Exactly, and this is a clear correlation which has been demonstrated multiple times over the past few months. WCG should put ARP on standby for the moment, to maintain the over-all system health. EDIT: I will add that once a download connection is successful, the download speed is now good. The previous change to limit the number of concurrent connections appears to have worked in that regard.  “Where an excess of power prevails, property of no sort is duly respected. No man is safe in his opinions, his person, his faculties, or his possessions.” – James Madison (1792) [Edit 1 times, last edit by Paul Schlaffer at Dec 10, 2022 5:51:42 PM] |
||
|
|
Unixchick
Veteran Cruncher Joined: Apr 16, 2020 Post Count: 998 Status: Offline Project Badges: 
|
As a ARP fan, I have to say it is obvious that ARP is causing the download errors. Instead of stopping it, I would suggest slowing down ARP. release a small amount of ARP in bursts. Then the download errors would be intermittent and not constant. This would also give the WCG moments to test if they have a solution.
Since ARP is the only project that has an end date, it would be good to keep it moving, even if much slower. |
||
|
|
Taurus Oldbull
Advanced Cruncher US Joined: Nov 26, 2020 Post Count: 53 Status: Offline Project Badges: 
|
Hi Cyclops,
Please note that all the ARP tasks I received have 3-day deadlines, instead of the normal 5-day deadlines, even though they all have -0 or -1 at the end of their identifiers, which should indicate that they are initial runs, not repeats. Thanks. |
||
|
|
PMH_UK
Veteran Cruncher UK Joined: Apr 26, 2007 Post Count: 774 Status: Recently Active Project Badges: 
|
Cyclops,
----------------------------------------There is a post on WCG facebook page under the 22 Nov challenge post that should be sanitised and responded to. Paul.
Paul.
|
||
|
|
Paul Schlaffer
Senior Cruncher USA Joined: Jun 12, 2005 Post Count: 247 Status: Offline Project Badges:
|
Hi Cyclops, Please note that all the ARP tasks I received have 3-day deadlines, instead of the normal 5-day deadlines, even though they all have -0 or -1 at the end of their identifiers, which should indicate that they are initial runs, not repeats. Thanks. FYI, Those that I've received do have the “Where an excess of power prevails, property of no sort is duly respected. No man is safe in his opinions, his person, his faculties, or his possessions.” – James Madison (1792) [Edit 1 times, last edit by Paul Schlaffer at Dec 10, 2022 6:02:41 PM] |
||
|
|
Speedy51
Veteran Cruncher New Zealand Joined: Nov 4, 2005 Post Count: 1297 Status: Offline Project Badges: 
|
As a ARP fan, I have to say it is obvious that ARP is causing the download errors. Instead of stopping it, I would suggest slowing down ARP. release a small amount of ARP in bursts. Then the download errors would be intermittent and not constant. This would also give the WCG moments to test if they have a solution. Since ARP is the only project that has an end date, it would be good to keep it moving, even if much slower. I agree keep the project going.  [Edit 1 times, last edit by Speedy51 at Dec 11, 2022 8:01:00 AM] |
||
|
|
Paul Schlaffer
Senior Cruncher USA Joined: Jun 12, 2005 Post Count: 247 Status: Offline Project Badges:
|
As a ARP fan, I have to say it is obvious that ARP is causing the download errors. Instead of stopping it, I would suggest slowing down ARP. release a small amount of ARP in bursts. Then the download errors would be intermittent and not constant. This would also give the WCG moments to test if they have a solution. Since ARP is the only project that has an end date, it would be good to keep it moving, even if much slower. I agree keep the project going. IMHO you are not going to achieve anything with faster storage if the Internet connection cannot handle the load on the download servers. They stated in a prior update that the download issue wasn't due to the internet connection speed/capacity. This is the reason they're adding a higher performance SSD storage layer. “Where an excess of power prevails, property of no sort is duly respected. No man is safe in his opinions, his person, his faculties, or his possessions.” – James Madison (1792) [Edit 4 times, last edit by Paul Schlaffer at Dec 10, 2022 10:35:16 PM] |
||
|
|
Unixchick
Veteran Cruncher Joined: Apr 16, 2020 Post Count: 998 Status: Offline Project Badges:
|
Hi Cyclops, Please note that all the ARP tasks I received have 3-day deadlines, instead of the normal 5-day deadlines, even though they all have -0 or -1 at the end of their identifiers, which should indicate that they are initial runs, not repeats. Thanks. Taurus (and Cyclops), The shorter deadline could be due to an "extreme" lagard ARP. Since each iteration is built on the previous iteration, then if a particular parcel falls behind enough a shorter deadline was given. So I'm guessing the iteration number in the WU name is low . ARP1_parcelNumber_iterationNumber_WUcopy. I'll take a guess and say it is less than 120. |
||
|
|
Speedy51
Veteran Cruncher New Zealand Joined: Nov 4, 2005 Post Count: 1297 Status: Offline Project Badges:
|
[ They stated in a prior update that the download issue wasn't due to the internet connection speed/capacity. This is the reason they're adding a higher performance SSD storage layer. Thank you for pointing this out. I have removed what I said in my previous post ---------------------------------------- |
||
|
|
|