Hours not being added, last day has negative values.
Server status shows traitement_v4 rise3 Not Running and credit disabled.

---

Paul H.

Hours appear normal now.
Thanks.

---

Paul H.

Hours not being added, last day has negative values.
Server status shows traitement_v4 rise3 Not Running and credit disabled.

What Project is this?

What Project is this?

He was talking about THIS project - WuProp.
Hours go negative at times on our status when things are not working right.

Hours are not being added in the past day for some reason. I noticed some tasks have been running for nearly a day. Some are at 23/25 hours as of this writing.

I also noticed a new test app that came up yesterday. Seb, what has changed? Will these tasks complete or do I have to abort and lose and entire day of hours to get another task?

Edit:
I missed it at first but the Linux app was also updated yesterday to v4.36. What triggered me to look was when a new exe downloaded per the log file after suspending/resuming the task:

105426 WUProp@Home 9/7/2026 8:47:11 AM task data_collect_v4_1784690102_1988672_0 suspended by user
105427 WUProp@Home 9/7/2026 8:47:15 AM task data_collect_v4_1784690102_1988672_0 resumed by user
105428 WUProp@Home 9/7/2026 8:47:19 AM Computation for task data_collect_v4_1784690102_1988672_0 finished
105429 WUProp@Home 9/7/2026 8:47:21 AM Started upload of data_collect_v4_1784690102_1988672_0_r1223034273_0
105430 WUProp@Home 9/7/2026 8:47:23 AM update requested by user
105431 WUProp@Home 9/7/2026 8:47:24 AM Finished upload of data_collect_v4_1784690102_1988672_0_r1223034273_0 (238351 bytes)
105432 WUProp@Home 9/7/2026 8:47:28 AM Sending scheduler request: Requested by user.
105433 WUProp@Home 9/7/2026 8:47:28 AM Reporting 1 completed tasks
105434 WUProp@Home 9/7/2026 8:47:28 AM Requesting new tasks for CPU
105435 WUProp@Home 9/7/2026 8:47:30 AM Scheduler request completed: got 1 new tasks
105436 WUProp@Home 9/7/2026 8:47:30 AM Project requested delay of 7 seconds
105437 WUProp@Home 9/7/2026 8:47:32 AM Started download of data_collect_v4_436_x86_64-pc-linux-gnu__nci
105438 WUProp@Home 9/7/2026 8:47:32 AM Started download of data_collect_v4_1788697069_29324
105439 WUProp@Home 9/7/2026 8:47:33 AM Finished download of data_collect_v4_1788697069_29324 (63 bytes)

This version does not appear to run for 1 hour and stop but appear to complete at some random time, mostly between 1 and 2 hours.

Two of my tasks completed after suspending/resuming that were at like 80k seconds. A third would not complete at 3.5 hours and was aborted.

What is different about 4.36? It does not appear to be an improvement.

Edit2:
After the two complete day tasks completed, no more hours were updated. 32 threads of PrimeGrid work were running the entire time but no hours were added. Did the way WUprop check for running work change, it think nothing was running so the WUProp task just kept running?

Tasks that ran for longer than 1 hour have this in the task log even though
11:29:31 (719809): Activite reseau suspendue : -1
The ones that ran for a day have hundreds of lines. The ones that run for an hour, none. The ones that ran for more than an hour, have some.

I have not suspended network activity or lost it. The PC that had the task run for a day are on WiFi. They are the client running CPU work. The client running GPUs on the same PCs did not run for a day.

The new app is broken. Why is network activity now a factor and exclude valid data? What about people that only have a hotspot on their phone?

v4.36 continues to miss thousands of hours a day. Even hours on wired connections that never lost network connectivity. Even if they did, BOINC is not affected.

Same here, noticed longer run times. Have 1 at 13 Hr's right now ... So far all the longer run Wu's have reported, if they reported the Hr's for the longer run time I don't know.

---

https://signature.statseb.fr/sig-1323.png

I can confirm same problem with hours not being reported by the WuProp 4.36 version on Linux.

It seems hit and miss as to if the tasks will run for the usual 1 hour or something different.

All the 1+ hour tasks have this in the Stderr output: Activite reseau suspendue : -1

Here is the Stderr output of a v4.36 task that took over 1 hour to complete:

<core_client_version>8.2.5</core_client_version>
<![CDATA[
<stderr_txt>
09:36:21 (1029252): initialisation
Erreur assignation taille wu
Erreur assignation taille application
09:43:21 (1029252): Erreur reception active_result
10:36:21 (1029252): Activite reseau suspendue : -1
10:37:21 (1029252): Activite reseau suspendue : -1
10:38:21 (1029252): Activite reseau suspendue : -1
10:39:21 (1029252): Activite reseau suspendue : -1
10:40:21 (1029252): Activite reseau suspendue : -1
10:41:21 (1029252): Activite reseau suspendue : -1
10:42:21 (1029252): Activite reseau suspendue : -1
10:43:21 (1029252): Activite reseau suspendue : -1
10:44:21 (1029252): Activite reseau suspendue : -1
10:45:21 (1029252): Activite reseau suspendue : -1
10:46:21 (1029252): Activite reseau suspendue : -1
10:47:21 (1029252): Activite reseau suspendue : -1
10:48:21 (1029252): Activite reseau suspendue : -1
10:49:21 (1029252): Activite reseau suspendue : -1
10:50:21 (1029252): Activite reseau suspendue : -1
10:51:21 (1029252): Activite reseau suspendue : -1
10:52:21 (1029252): Activite reseau suspendue : -1
10:53:21 (1029252): Activite reseau suspendue : -1
10:54:21 (1029252): Activite reseau suspendue : -1
10:55:21 (1029252): Activite reseau suspendue : -1
10:56:21 (1029252): Activite reseau suspendue : -1
10:57:21 (1029252): Activite reseau suspendue : -1
10:58:21 (1029252): Activite reseau suspendue : -1
10:59:21 (1029252): Activite reseau suspendue : -1
11:00:21 (1029252): Activite reseau suspendue : -1
11:01:21 (1029252): Activite reseau suspendue : -1
11:02:21 (1029252): Activite reseau suspendue : -1
11:03:21 (1029252): Activite reseau suspendue : -1
11:04:21 (1029252): Activite reseau suspendue : -1
11:05:21 (1029252): Activite reseau suspendue : -1
Erreur assignation taille wu
Erreur assignation taille application
11:06:21 (1029252): called boinc_finish(0)

</stderr_txt>
]]>

Here is the Stderr output of a v4.36 task that took 1 hour to complete:

<core_client_version>8.2.5</core_client_version>
<![CDATA[
<stderr_txt>
12:06:53 (1107107): initialisation
Erreur assignation taille wu
Erreur assignation taille application
Erreur assignation taille wu
Erreur assignation taille application
13:06:53 (1107107): called boinc_finish(0)

</stderr_txt>
]]>

Same here, I only run Linux.

Version went from
4.33 at 6 Sep 2026, 11:51:17 UTC
4.34 at 6 Sep 2026, 12:25:02 UTC (execv: Exec format error)
4.35 at 6 Sep 2026, 12:51:33 UTC (OK)
4.36 at 6 Sep 2026, 13:25:20 UTC (some OK, others got messages but ended, a few just keep going)

Tried brief suspend/resume but still ran, aborted.
Also tried reset but still some errors.

Edit:
Applications shows updates for Linux & mac only.
Test app updated at 4.39 6 Sep 2026, 10:02:21 UTC, server status shows 10000 unsent, 0 in progress.

---

Paul H.

I just had one cancelled by server.....

Replaced by version 4.37

Two 4.37 units done, both valid but 1 had:
"18:10:38 (997175): Erreur assignation user_friendly_name (node app) - app_name: "

---

Paul H.

My Linux box first received a 4.37 version on it's last contact about 50 minutes ago. We will see how that works out.

There's two Linux apps currently available.
* One is a beta test:
Linux running on an AMD x86_64 or Intel EM64T CPU 4.39 (nci) 6 Sep 2026, 10:02:21 UTC

* The other is the current '"main / stable" version:
Linux running on an AMD x86_64 or Intel EM64T CPU 4.37 (nci) 8 Sep 2026, 15:33:47 UTC

So far my Linux box completed 2 tasks under v4.37 lasting right about 1 hour.

But the total for the last 24 "wall" hours is still stuck on 35.70 hours out of 46 to 48 hours expected.[/quote]

Same here, I only run Linux.

Version went from
4.33 at 6 Sep 2026, 11:51:17 UTC
4.34 at 6 Sep 2026, 12:25:02 UTC (execv: Exec format error)
4.35 at 6 Sep 2026, 12:51:33 UTC (OK)
4.36 at 6 Sep 2026, 13:25:20 UTC (some OK, others got messages but ended, a few just keep going)

Tried brief suspend/resume but still ran, aborted.
Also tried reset but still some errors.

Edit:
Applications shows updates for Linux & mac only.
Test app updated at 4.39 6 Sep 2026, 10:02:21 UTC, server status shows 10000 unsent, 0 in progress.

I think the suspend/resume only worked with 4.36 if the task thought it had ran/collected data for 1 hour. But for some reason 4.36 stopped collecting data when it erroneously thought networking was off.

Good to see things have resumed.

Appears fixed.

---

Paul H.

NFS cuda-sieve is recorded in a weird way:
It isn't shown as activity over the last 24h, it isn't shown in the machine's daily activity, but is is shown as around 99% of the machine's weekly activity.
As I just now looked at the machine's activity, it didn't occur to me to look at the totals, thus I haven't been able to confirm wether or not the values have been added to the total hours for this application. Will keep an eye on it.

---

- - - - - - - - - -
Greetings, Jens

It's funny. Since clicking the weekly activity the percentage has gone down although the machine is still doing the work.
Overall cuda-sieve time kept to what it was before.
I'm wondering if I'd get blocked from WUProp if only running NFS cuda-sieve and WUProp. ^^

Btw.: cuda-sieve for moderate numbers is counting.

---

- - - - - - - - - -
Greetings, Jens

I am also seeing no hours for the "cuda-sieve" app, even though I have two machines crunching it 24/7, with two tasks each. This just started (not) happening in the past day or two, I think. FWIW, the other NFS CUDA app is being counted correctly.

Looks fixed now.
Great!
Thx.

---

- - - - - - - - - -
Greetings, Jens

There was a plan_class change to the cudasieve app. Maybe that cause the issue?

If it was, then the cuda moderate app also updated to an 8030 to 5030 plan_class.