I've have many Postponed: VM job unmanageable, restarting later WUs on my Linux., seems to getting more worse lately.

I'm shutting my PC down for the night so when I startup the next day, the postponed tasks just continue. I'm running 9 python and last night I had 5 tasks postponed, shortly after starting up today I already have 2 tasks postponed.

This is very annoying!

Then I'm not alone with the problem. Wonder if this is limited to Linux though.
I'm have no major problem with postponed work... but halting and stop reviving new work is more annoying.

I wonder if this works (does not last between reboots though if understood correctly. Unsure if adding the timer as sudo makes it run though)... I could live with daily automatic restarts.

sudo systemd-run --OnCalendar 'daily' systemctl restart boinc-client

https://linuxconfig.org/how-to-schedule-tasks-with-systemd-timers-in-linux

Then I'm not alone with the problem. Wonder if this is limited to Linux though.
I'm have no major problem with postponed work... but halting and stop reviving new work is more annoying.

I wonder if this works (does not last between reboots though if understood correctly. Unsure if adding the timer as sudo makes it run though)... I could live with daily automatic restarts.

sudo systemd-run --OnCalendar 'daily' systemctl restart boinc-client

https://linuxconfig.org/how-to-schedule-tasks-with-systemd-timers-in-linux

Correction

sudo systemd-run --on-calendar 'daily' systemctl restart boinc-client

The issue I have with restarting BOINC to get the postponed to run, is that you lose time on all running tasks from all projects, due to the checkpoints.

I've have many Postponed: VM job unmanageable, restarting later WUs on my Linux., seems to getting more worse lately.

I'm shutting my PC down for the night so when I startup the next day, the postponed tasks just continue. I'm running 9 python and last night I had 5 tasks postponed, shortly after starting up today I already have 2 tasks postponed.

This is very annoying!

you might just do a project max_concurrent setting then. get it down to the number of tasks you can run without problems. But are you still doing Einstein and Milkyway?


Postponed: VM job unmanageable - I had that a lot on QuChem and had to quit.
I guess it wasn't good at sharing with other VM projects.
Maybe you should isolate RAH pythons to a machine that has no other VM work and see if the error still pops up.

What computers are doing what? You have 3. You need to talk about which one.

According to Boinc logs, the work should restart by itself, 86400 (1 day later). But if it fails over and over? Automatic removal by deadline? So far haven't seen a halted one actually finish when restarted. It halts at the same percentage. BUT at least the reboot is fast enough to work as a workaround to allow downloading of new work.

I've have many Postponed: VM job unmanageable, restarting later WUs on my Linux., seems to getting more worse lately.

I'm shutting my PC down for the night so when I startup the next day, the postponed tasks just continue. I'm running 9 python and last night I had 5 tasks postponed, shortly after starting up today I already have 2 tasks postponed.

This is very annoying!

you might just do a project max_concurrent setting then. get it down to the number of tasks you can run without problems. But are you still doing Einstein and Milkyway?


Postponed: VM job unmanageable - I had that a lot on QuChem and had to quit.
I guess it wasn't good at sharing with other VM projects.
Maybe you should isolate RAH pythons to a machine that has no other VM work and see if the error still pops up.

What computers are doing what? You have 3. You need to talk about which one.

In my case there are no other jobs than Rosetta. Nothing but Rosetta uses Vbox. Many work finish and uploads, but the ones that fail always seem to fail. 3 postponed right now. All 3 begins with "aagb". Does those letters mean anything that could perhaps explain it?

Anyone running python jobs on Linux?
Currently without work due to WCG moving and I more or less stopped using Rosetta for some time due to no work every now and then. But now back and figured I'd try these now jobs out.
I setup virtualbox and work is coming in and grabbing a pretty decent amount of memory ;-) All seems fine... However, after some time (it may even succeed in finishing some jobs) it ends up as "Postponed: VM job unmanageable, restarting later". This happens on my second experimental semi-broken laptop too. No new tasks is started either.
In this case running Manjaro. Rosetta or WCG jobs have never been a issue before. It's Vbox jobs that is the issue.
Something seems to prevent further use of Vbox.
Atleast I know it's not a ram issue. Got roughly 16GB to spare on main computer. Dynamically adjusted swap. Most often 0 or close to 0.

That's not a Linux problem, I get the same on Windows. I can't remember what I did to stop it, if I even did. I can't even remember if it was Rosetta or LHC or Cosmology.

A workaround is to close the Boinc client and reopen it (you don't need to restart the computer or VB itself). You need to properly exit Boinc Manager from the system tray (or whatever the Linux equivalent is), or exit using it's file menu, and tell it to close tasks when exiting. Give it a while to shut everything down (watch processor and disk usage to make sure), then start it again. This will reset the 1 day timer on retrying those tasks.

You can also try cleaning out the VB environment - although usually you would have an error in Boinc telling you it needs cleaning up. Open the VB manager and locate which jobs correspond to the tasks in Boinc, then delete the rest. They will usually be marked "inaccessible". Delete those.

I might have fixed it by changing the VB version. I know Cosmology hates version 6 so I changed to 5, but I thought Rosetta and LHC worked ok on both. If you change version, make sure you install the extensions for that version.

And make sure you have VB extensions installed, not sure if rosetta uses them, but best to have them there.

Yes...aagb seems to be a task group that gets stuck.

I posted about that in number crunching as did someone else.

Can you assist a totally clueless brother cruncher?

I've looked everywhere I can think of and cannot find that option


Where is it ?

Thanks

Go to your account homepage on the Rosetta website, click on "View" next to "Computers on this account", then click on "Details".

At the bottom, there is a toggle switch headed "VirtualBox VM jobs". If it says "Allow", click it, and you will start getting Pythons once your Boinc Manager communicates with the server.

If you click "Allow" and then "Return to host page", you will see that the toggle now says "Skip" (and clicking it would then stop you getting Pythons).

Well, that was easy

Thanks

Well, that was easy

Thanks

RAH guys like to hide things, so this is a common question.
It's probably logical to put it there, but you would think it would be in project computing preferences.

If you run into trouble completing the task, come over to number crunching. We discuss all the problems there. If you get a task that stalls and it has the first letters of its name being aagb, just abort it.
Windows users are having troubles with these tasks.

I was notified by Einstein@home that there are no projects left for my Rpi 4's. Now I see there are no projects left for them here as I started getting "more memory needed" messages recently when attempting to poll updates, and then confirm with the messages I read here. I made sure to get 3 of them, all 4Gb (the 8Gb was not out when I purchased them), mainly for the use for Rosetta and migrated most of them to add Einstein when projects were low here.

I ran the Pi's because the electricity is negligible and I felt it was for a good cause. Now C19 is rather under control, I see the needs of science have outlived my equipment and they want me to dedicate my GPU's or larger equipment to the cause, using my electricity that makes an impact. Well sorry, those get reserved for crypto, see those when used for that earn $$$ to pay for their use for that. So no hard feelings, I feel that it was a good run but until needed again my 3 lonely little Pi's are shutdown and being moved to other more fruitful projects.

Good luck on future endeavors, email us if you have new projects that can include the ARM group, but until then if it is GPU power you need then I suggest you use the power at hand and grant money and get the super computer that you are so desiring your projects toward, or put it toward lease time. To ignore the smaller parts of the whole is basically a defeat of the goals set forth by the project when it started to begin with, to use any an all resources to band together and solve a common task, not pick and choose which are easier or faster to get the job done.

Until next time........

I was notified by Einstein@home that there are no projects left for my Rpi 4's. Now I see there are no projects left for them here as I started getting "more memory needed" messages recently when attempting to poll updates, and then confirm with the messages I read here. I made sure to get 3 of them, all 4Gb (the 8Gb was not out when I purchased them), mainly for the use for Rosetta and migrated most of them to add Einstein when projects were low here.

I ran the Pi's because the electricity is negligible and I felt it was for a good cause. Now C19 is rather under control, I see the needs of science have outlived my equipment and they want me to dedicate my GPU's or larger equipment to the cause, using my electricity that makes an impact. Well sorry, those get reserved for crypto, see those when used for that earn $$$ to pay for their use for that. So no hard feelings, I feel that it was a good run but until needed again my 3 lonely little Pi's are shutdown and being moved to other more fruitful projects.

Good luck on future endeavors, email us if you have new projects that can include the ARM group, but until then if it is GPU power you need then I suggest you use the power at hand and grant money and get the super computer that you are so desiring your projects toward, or put it toward lease time. To ignore the smaller parts of the whole is basically a defeat of the goals set forth by the project when it started to begin with, to use any an all resources to band together and solve a common task, not pick and choose which are easier or faster to get the job done.

Until next time........

Einstein@Home is handing out Radio Wave tasks to my Android phones, do those not run on Pis? They did have a problem recently where they stopped being handed out for about a week, but they are back.

I've tried crypto, it doesn't even pay for the electricity. You could try Gridcoin, that means you can do work for Boinc and get cash at the same time.

I was notified by Einstein@home that there are no projects left for my Rpi 4's. Now I see there are no projects left for them here as I started getting "more memory needed" messages recently when attempting to poll updates, and then confirm with the messages I read here. I made sure to get 3 of them, all 4Gb (the 8Gb was not out when I purchased them), mainly for the use for Rosetta and migrated most of them to add Einstein when projects were low here.

I ran the Pi's because the electricity is negligible and I felt it was for a good cause. Now C19 is rather under control, I see the needs of science have outlived my equipment and they want me to dedicate my GPU's or larger equipment to the cause, using my electricity that makes an impact. Well sorry, those get reserved for crypto, see those when used for that earn $$$ to pay for their use for that. So no hard feelings, I feel that it was a good run but until needed again my 3 lonely little Pi's are shutdown and being moved to other more fruitful projects.

Good luck on future endeavors, email us if you have new projects that can include the ARM group, but until then if it is GPU power you need then I suggest you use the power at hand and grant money and get the super computer that you are so desiring your projects toward, or put it toward lease time. To ignore the smaller parts of the whole is basically a defeat of the goals set forth by the project when it started to begin with, to use any an all resources to band together and solve a common task, not pick and choose which are easier or faster to get the job done.

Until next time........

Einstein@Home is handing out Radio Wave tasks to my Android phones, do those not run on Pis? They did have a problem recently where they stopped being handed out for about a week, but they are back.

I've tried crypto, it doesn't even pay for the electricity. You could try Gridcoin, that means you can do work for Boinc and get cash at the same time.

Gridcoin is worth: $0.009278. (9/100th of a cent) that is not even a realistic price. And that they want to control your systems. No thanks. My system is used for what I want and to occupy my system with 9/100th of a cent, its not even a remote thought.

Gridcoin is worth: $0.009278. (9/100th of a cent) that is not even a realistic price. And that they want to control your systems. No thanks. My system is used for what I want and to occupy my system with 9/100th of a cent, its not even a remote thought.

I assume you're talking about using a pool. I just gave up on that idea, I couldn't for example turn Python on and off. Instead I now use Gridcoin directly from my wallet. It doesn't control anything. I get paid for any of the Boinc projects that are on their whitelist. The ones not on it I'm free to still use but I get no coins from them. It doesn't pay for the electricity, but I'm paid enough to refund me for my hardware in 1.5 years. So I might aswell take it.

Gridcoin is worth: $0.009278. (9/100th of a cent) that is not even a realistic price. And that they want to control your systems. No thanks. My system is used for what I want and to occupy my system with 9/100th of a cent, its not even a remote thought.

I assume you're talking about using a pool. I just gave up on that idea, I couldn't for example turn Python on and off. Instead I now use Gridcoin directly from my wallet. It doesn't control anything. I get paid for any of the Boinc projects that are on their whitelist. The ones not on it I'm free to still use but I get no coins from them. It doesn't pay for the electricity, but I'm paid enough to refund me for my hardware in 1.5 years. So I might aswell take it.

1.5 years just for the hardware? Well ok.

Gridcoin is worth: $0.009278. (9/100th of a cent) that is not even a realistic price. And that they want to control your systems. No thanks. My system is used for what I want and to occupy my system with 9/100th of a cent, its not even a remote thought.

I assume you're talking about using a pool. I just gave up on that idea, I couldn't for example turn Python on and off. Instead I now use Gridcoin directly from my wallet. It doesn't control anything. I get paid for any of the Boinc projects that are on their whitelist. The ones not on it I'm free to still use but I get no coins from them. It doesn't pay for the electricity, but I'm paid enough to refund me for my hardware in 1.5 years. So I might aswell take it.

1.5 years just for the hardware? Well ok.

Since hardware lasts for more than 1.5 years, that means I'm getting free hardware and then some. And once you've set up Gridcoin, there's no work involved, and I can use Boinc as normal and change settings as I normally would, and get stats as I normally would. Might aswell take a free lunch.

You and your research will have our full support