Received 64 ARP1-tasks. Despite 215 times the message "transient HTTP error" it took only three hours to download them thanks to setting max_file_xfers to 1 in cc_config.xml.

They are all generations 126 and 127. The deadline is 6 days. WCG (IBM) has experimented with this value in the past. In October 2019 it started with 7 days, in July 2021 this changed to 8 days, in September 2021 there was a deadline of 4½ days for a short while which changed back to 8 days. Only in February 2022, just before the 'migration', the deadline became 6 days for a while, then shortly even 3 days to clean them all up.

EDIT: Just received three more ARP1s with a deadline of only 3 days in generation 125:

 <7>   ARP1_0025619_125_0  Linux         In Progress           2022-07-16T12:29:20  2022-07-19T12:29:20
 <7> * ARP1_0025619_125_1  Fedora Linux  In Progress           2022-07-16T12:31:33  2022-07-19T12:31:33

 <8>   ARP1_0007468_125_0  Linux Pop     In Progress           2022-07-16T12:30:46  2022-07-19T12:30:46
 <8> * ARP1_0007468_125_1  Fedora Linux  In Progress           2022-07-16T12:29:25  2022-07-19T12:29:25

 <9>   ARP1_0026015_125_0  Linux Gentoo  In Progress           2022-07-16T12:32:04  2022-07-19T12:32:04
 <9> * ARP1_0026015_125_1  Fedora Linux  In Progress           2022-07-16T12:29:25  2022-07-19T12:29:25

Unixchick & Mark,

Access to those links seems to suffer from the same problems that are causing the "transient HTTP error" issues on downloads (503 - down/overloaded?). However, I finally managed to access all three and have taken a snapshot...

I must admit it looks as if the data presented is for the last day of IBM's WCG, but I'll use my snapshots to check again tomorrow and subsequently. If it turns out they don't update, perhaps sorting that out can be added to the "snagging list"...

Cheers - Al.

Al,
Thank you for looking into this. All 3 links are now giving results for me.
As I understand it, there is a problem with statistics from work units crunched recently not appearing on the website. I am hoping that the data presented by these links will start to update once that problem has been solved. But that may be over-optimistic!
Cheers,
Mark

The two "threads" in my two posts above (regarding these statistics, and assessing "extreme" and "accelerated" work unit candidates) tie together on a point which may well make your prediction valid... If IBM gave Krembil a proper, fully documented, briefing on all the oddities associated with ARP1 then dealing with future short-limit work units shouldn't present any problems. And, of course, if the mechanisms and data structures are still present, making new scripts to provide the three simple status files would be trivial.

(The alternative would presumably be a reversion to just running ARP1 tasks "as is" with the almost certain return of stuck units and unnecessary delays to the completion of the project. With enough clued-in volunteers, it might be that the community could spot these issues, but it would be better done "in house"...)

However, some folks seem to doubt the likelihood of a well-documented hand-over -- only Krembil and IBM know the truth of that... So you could be right :-(

Cheers - Al.

Nice, Robert! You have dug up one that has been reported by knreed as restarted from the beginning.

Adri