Posts by shanen

61) Message boards : Number crunching : Recovered the bug report, but... (Message 88373)
Posted 24 Feb 2018 by Profile shanen
Post:
Okay, the first half made it... Now let's try the second half...

...do anything? I went off muttering to myself.

This morning the stuck task is running as though nothing funny has been going on. The Progress is now over 80% and the Remaining is showing an apparently normal value over 1 hour. Most interestingly, the 28 hours of the Elapsed time has been reset down to some 7 hours, and it is counting up.

Apparently nothing to see there? Just buggy software.

Just checked a Windows 10 box. I wanted to shut it down last night for my wife's convenience this morning, but it had two tasks with over two hours of elapsed time since the last checkpoint... Buggy, buggy software.
62) Message boards : Number crunching : Recovered the bug report, but... (Message 88372)
Posted 24 Feb 2018 by Profile shanen
Post:
Nope, the long report did not make it here. Some kind of illegal characters in that report?

Let me try a minor tweak of the bizarre sort?

Update on that sick task, though mostly I don't want to spend more time than needed "supporting" any BOINC project with my actual thinking time...

The stuck task accumulated over 28 hours of running time and was obviously making no progress. I should have noticed that the "Remaining" field was blanked as though it had already finished, but I kept staring at the increasing time in the "Elapsed" column and the stuck percentage in the "Progress" column. (Another annoyance is the different ordering of those columns on various machines, and that one is a Linux box with the "Remaining" column in an especially less prominent place.)

So I suspended that task and let another one start running. So far, so good, just as expected. Then I resumed the long task, and even though it was getting close to its deadline, nothing happened. Not as expected. The imminent deadline task should restart right away. Apparently the BOINC Manager regarded it as finished after all? But it did not upload or do anything? I went off muttering to myself.
63) Message boards : Number crunching : Please stop sending tasks with 3-day deadlines (Message 88371)
Posted 24 Feb 2018 by Profile shanen
Post:
Update on that sick task, though mostly I don't want to spend more time than needed "supporting" any BOINC project with my actual thinking time...

The stuck task accumulated over 28 hours of running time and was obviously making no progress. I should have noticed that the "Remaining" field was blanked as though it had already finished, but I kept staring at the increasing time in the "Elapsed" column and the stuck percentage in the "Progress" column. (Another annoyance is the different ordering of those columns on various machines, and that one is a Linux box with the %
64) Message boards : Number crunching : Please stop sending tasks with 3-day deadlines (Message 88370)
Posted 24 Feb 2018 by Profile shanen
Post:
Update on that sick task, though mostly I don't want to spend more time than needed "supporting" any BOINC project with my actual thinking time...

The stuck task accumulated over 28 hours of running time and was obviously making no progress. I should have noticed that the "Remaining" field was blanked as though it had already finished, but I kept staring at the increasing time in the "Elapsed" column and the stuck percentage in the "Progress" column. (Another annoyance is the different ordering of those columns on various machines, and that one is a Linux box with the %
65) Message boards : Number crunching : More Rosetta@home bugs (Message 88369)
Posted 24 Feb 2018 by Profile shanen
Post:
If at third try you don't succeed, change your direction and try again. Here's the longer description of the bug I was trying to report in the other thread (when the website bug blocked it). I figured out a third way to recover it:

Update on that sick task, though mostly I don't want to spend more time than needed "supporting" any BOINC project with my actual thinking time...

The stuck task accumulated over 28 hours of running time and was obviously making no progress. I should have noticed that the "Remaining" field was blanked as though it had already finished, but I kept star
66) Message boards : Number crunching : Recovered the bug report, but... (Message 88368)
Posted 24 Feb 2018 by Profile shanen
Post:
Hmm... So now I've been able to post several times as long as I don't include the full report of the original problem. Some computers are like that?

Let me post this reply and then add the long report as a reply to this one?
67) Message boards : Number crunching : More Rosetta@home bugs (Message 88367)
Posted 24 Feb 2018 by Profile shanen
Post:
If at third try you don't succeed, change your direction and try again. Here's the longer description of the bug I was trying to report in the other thread (when the website bug blocked it). I figured out a third way to recover it:

Update on that sick task, though mostly I don't want to spend more time than needed "supporting" any BOINC project with my actual thinking time...

The stuck task accumulated over 28 hours of running time and was obviously making no progress. I should have noticed that the "Remaining" field was blanked as though it had already finished, but I kept star
68) Message boards : Number crunching : Recovered the bug report, but... (Message 88366)
Posted 24 Feb 2018 by Profile shanen
Post:
For reasons unknown, the website keeps rejecting the full report. I do have a copy here, but when I try to publish it on the Rosetta website, the connection gets broken for reasons unknown apart from very buggy software.
69) Message boards : Number crunching : Please stop sending tasks with 3-day deadlines (Message 88365)
Posted 24 Feb 2018 by Profile shanen
Post:
Update on that sick task, though mostly I don't want to spend more time than needed "supporting" any BOINC project with my actual thinking time...

The stuck task accumulated over 28 hours of running time and was obviously making no progress. I should have noticed that the "Remaining" field was blanked as though it had already finished, but I kept staring at the increasing time in the "Elapsed" column and the stuck percentage in the "Progress" column. (Another annoyance is the different ordering of those columns on various machines, and that one is a Linux box with the %
70) Message boards : Number crunching : Please stop sending tasks with 3-day deadlines (Message 88364)
Posted 24 Feb 2018 by Profile shanen
Post:
Probably not a useful diagnostic, but why is it turning off my sig as I try to dispose of this last stutter?
71) Message boards : Number crunching : Please stop sending tasks with 3-day deadlines (Message 88363)
Posted 24 Feb 2018 by Profile shanen
Post:
Apparently there were 4 stutters on the failure to post. Underlying that there was a partial autosave? But now I can't find any easy way to delete the stutters except overwrite them more briefly.
72) Message boards : Number crunching : Please stop sending tasks with 3-day deadlines (Message 88362)
Posted 24 Feb 2018 by Profile shanen
Post:
Update on that sick task, though mostly I don't want to spend more time than needed "supporting" any BOINC project with my actual thinking time...

The stuck task accumulated over 28 hours of running time and was obviously making no progress. I should have noticed that the "Remaining" field was blanked as though it had already finished, but I kept staring at the increasing time in the "Elapsed" column and the stuck percentage in the "Progress" column. (Another annoyance is the different ordering of those columns on various machines, and that one is a Linux box with the %

Pretty annoying. I typed much more about the bug and what happened to it. Then I noted another bug. Then I tried to save the completed post and the buggy website apparently crashed it out, though it double posted parts of my attempts?

What started as a waste of time is thus extended into a longer waste of more time.

My conclusion in the original comment was buggy, buggy software. Now I have to upgrade that to buggy, buggy, buggy software.

I think the funniest thing is that I defensively saved the long comment to my clipboard before the first save attempt. Then at some point I put something else in the clipboard. *SIGH*
73) Message boards : Number crunching : Please stop sending tasks with 3-day deadlines (Message 88361)
Posted 24 Feb 2018 by Profile shanen
Post:
Update on that sick task, though mostly I don't want to spend more time than needed "supporting" any BOINC project with my actual thinking time...

The stuck task accumulated over 28 hours of running time and was obviously making no progress. I should have noticed that the "Remaining" field was blanked as though it had already finished, but I kept staring at the increasing time in the "Elapsed" column and the stuck percentage in the "Progress" column. (Another annoyance is the different ordering of those columns on various machines, and that one is a Linux box with the %
74) Message boards : Number crunching : Please stop sending tasks with 3-day deadlines (Message 88354)
Posted 23 Feb 2018 by Profile shanen
Post:
I'm just trying to suggest a minimal solution, but sometimes I don't know why I stick with this project. I've already quit running several others for lesser annoyances... The latest and current annoyance is an rb task that has over 20 hours of running time and less than 50% progress. Normal working time is supposed to be around 8 hours (though many run 12 hours with little extra credit earned).

I've also expressed my deeper concerns that such things make the project look buggy. If I were an actual working scientist, then I'd want any results from Rosetta@home to be double-checked. (Fortunately, I believe that most of the results are negative and there's little reason to double-check those.)
75) Message boards : Number crunching : Please stop sending tasks with 3-day deadlines (Message 88322)
Posted 20 Feb 2018 by Profile shanen
Post:
The largest annoyance with the Rosetta@home project is caused by 3-day tasks. Please stop sending them. (They weren't so bad when they were half-size, but they are seriously annoying now.)

The Mac problem seems to have been fixed several days ago.

This Message Board input form is slightly annoying because the input fields are quite hard to find. After you click into a field to give it the focus, then it gets shading that makes it more visible, but until that it's almost identical with the background.
76) Message boards : Number crunching : New MAC Boinc Client 7.8.6 (Message 88280)
Posted 13 Feb 2018 by Profile shanen
Post:
I think you mean that I should try checking on the option "Network activity always" rather than the option "Network activity based on preferences", so I'll give it a try. I cannot imagine why that would make a difference, and especially not when I manually click on Update from the Projects tab.

In my first test, the behavior seems unaffected and still incorrect. Normally (and the behavior remains normal as far as I can tell on all other computers), the Update click takes a few seconds and all of the pending units ("Ready to report") are cleared. Now it takes a long think, and usually nothing happens. The pending units are still sitting there, and it usually announces a long delay time, such as "Communications deferred 03:57:53" and starts counting down.

So far I don't believe that any of the tasks have missed their deadlines because of this peculiar behavior. It appears to succeed every day or three. It also seems to be keeping a sufficient number of tasks on hand so that the CPU is never idle.

Whatever is wrong, it is NOT specific to the new version, since the weird behavior started first and it then persuaded me to download and install the new version of the BOINC Manager.

The network configuration of the Mac is actually simple. It's always connected to my slower network. Most of my Ubuntu and one of the Windows boxen also use the slower network, but show no similar symptoms. I have other machines that sometimes use the faster network and sometimes use the slower one, and again none of them show any problems. And yes (as I believe I already noted in this discussion), I did try the Mac on the fast network without improvement.

Right now the Mac has 7 tasks in limbo. Seems pretty likely to me that there is something wrong on the Baker Lab end, and it affects Macs but not Windows 10 or Ubuntu Linux.
77) Message boards : Number crunching : New MAC Boinc Client 7.8.6 (Message 88260)
Posted 10 Feb 2018 by Profile shanen
Post:
Well, there is some other Mac-specific problem with Rosetta@home. Apparently unable to upload any finished tasks now. No problem on any other machines, but my MacBook Pro has a long list of "Ready to report" tasks there. Several days worth.

That did get me to try (per force) the upgrade to the new version. Went quite smoothly, but did not resolve the problem. I also tried an alternative network routing, which didn't help, but mostly I do NOT want to be bothered by such things and feel like I am wasting time when I fuss over the problems (which is especially annoying when the problems could be eliminated).
78) Message boards : Number crunching : New MAC Boinc Client 7.8.6 (Message 88250)
Posted 8 Feb 2018 by Profile shanen
Post:
I should have been more explicit. My concern is with the proper updating of the Mac BOINC Client. If it updates itself, then I know that everything has been handled properly, but there are various possible problems if I have to intervene and I don't even want to spend the time worrying about or studying those potential problems.
79) Message boards : Number crunching : New MAC Boinc Client 7.8.6 (Message 88224)
Posted 3 Feb 2018 by Profile shanen
Post:
I see my Mac is running an older version, but why isn't the update automated through the regular update system for Mac software? Doesn't seem to be worth the trouble unless maybe it fixes that old problem with the Notices tab, where it always returns the server not found error. (Actually, now I'm not certain if that error is only for the Mac version. The notices are not very frequent, but I'm sure that I haven't actually received any of them in a long time. Also, that's probably yet another BOINC-level problem out of the scope of Rosetta@home.)
80) Message boards : Number crunching : Progess going backwards? (Message 88065)
Posted 14 Jan 2018 by Profile shanen
Post:
I'm guessing you [Mod.Sense] mean "Leave non-GPU tasks in memory while suspended" on the "Disk and memory" tab. It is checked, but the task was NOT suspended when I noticed the drop in progress. I'm also hard pressed to imagine how it could have lost progress even if it had been suspended. The progress is somehow related to time rather than being a simple metric of work completed? The remaining is obviously related to time and I can see how suspension might confuse that one, but it's already an obviously flaky and nonlinear metric of whatever it's estimating.

Current annoyance is actually the checkpointing, especially on this machine. Whenever I want to shut it down, it seems like at least one of the active tasks has a large time since the last checkpoint. Right now I have a task that is almost half finished as it approaches 5 hours, but it's been more than 1-1/2 hours since the last checkpoint. (This one is a nRoCM... task, if that's worth knowing.) Can't suspend this machine because it's a cross-booter and I need the other OS sometimes.


Previous 20 · Next 20



©2024 University of Washington
https://www.bakerlab.org