Posts by Ardis

1) Questions and Answers : Windows : Error in request message: xp.get_tag() failed (Message 87859)
Posted 8 Dec 2017 by Ardis
Post:
Is it showing that it is reporting 1 completed task each time? Or does that vary? Do you still see a task in the this that is "ready to report"?


Yes, but it's the same task over and over. I tried unsuccessfully to abort it, then reset the whole project yesterday (lost one ~8 hour wu).

Right now there are 14 tasks uploading (10 are Rosetta), but they don't seem to be going anywhere. That used to be a pretty transient state, but not now. Thinking it might be a BOINC Manager issue, I got a new copy of 7.8.3 and ran the repair option. Didn't seem to make any difference.

Today there are a lot of "no command" error messages that look like this:

12/8/2017 12:39:10 AM | Rosetta@home | [error] Error reported by file upload server: no command


There are also several Transfers entries that seem to be sticking around for a long time too. The original 5 second error report cycle isn't happening any more, but we're still not acting very normal. Any Ideas?
2) Questions and Answers : Windows : Error in request message: xp.get_tag() failed (Message 87828)
Posted 5 Dec 2017 by Ardis
Post:
We don't seem to be feeling well. Rosetta is currently running three of four tasks, but every five seconds or so the BOINC Manager status says "Scheduler request in progress." The event log looks like this:

12/5/2017 1:15:06 AM | Rosetta@home | Sending scheduler request: To report completed tasks.
12/5/2017 1:15:06 AM | Rosetta@home | Reporting 1 completed tasks
12/5/2017 1:15:06 AM | Rosetta@home | Not requesting tasks: don't need (not highest priority project)
12/5/2017 1:15:07 AM | Rosetta@home | Scheduler request completed
12/5/2017 1:15:07 AM | Rosetta@home | Error in request message: xp.get_tag() failed

Then it's "communication deferred" for a few seconds, repeat ad nauseam.

Any ideas on how to improve our health?

Ardis
Win10, Mgr 7.8.3
3) Questions and Answers : Windows : CPU time (Message 20007)
Posted 10 Jul 2006 by Ardis
Post:
Marc,

Rosetta, and all BOINC projects, report CPU time, the amount of time your computer has spent actually crunching a work unit. The computer also needs to do other things to keep itself running, so CPU time on a project will always be less than the amount of "wall clock" time between starting and stopping. The most crunching I've seen here is 99%, so what you said is inevitable. Other factors also reduce CPU time on the project; even moving the mouse requires some calculation by your CPU that doesn't go toward crunching a WU. So anything you do with the computer will cut into crunch time, even though the clock on the wall is still ticking. The projects run at a really low priority, so think of it as unused CPU cycles. Another variable: BOINC allows you to choose general preferences that will further reduce the amount of CPU time spent on the projects.

Regards,

Ardis

Yeah, I know this is a late response Timliness is a virtue with which I have not been inflicted.
4) Message boards : Cafe Rosetta : Old Geezers Club (Message 19991)
Posted 10 Jul 2006 by Ardis
Post:
You folks bring back memories. Fortran, Cobol, yup. Still have the Kernighan & Ritchie C book. My first "Hello World" was on a *printing* terminal off of a Vax, running 300 baud....


That would be the one. With the operator precedence table on page 45. It was't until I read this quote from DMR himself that I finally understood the odd positioning of | & and ^. :)

http://www.quut.com/c/dmr-on-or.html


Ha! Hadn't seen that before, thanks for the reference. I do seem to remember a few lunch table discussions on operator precedence. "Why did they do it *that* way? In [my favorite language]...." My table is on page 49, so I looked in the front to check the edition (11th printing) and found this:

This book was set in Times Roman and Courier 12 by the authors, using a Graphic Systems phototypesetter driven by a PDP-11/70 running under the UNIX operating system.


Phototypesetter? Pretty revolutionary at the time.

Regards,

Ardis
5) Message boards : Cafe Rosetta : Old Geezers Club (Message 19673)
Posted 2 Jul 2006 by Ardis
Post:
You folks bring back memories. Fortran, Cobol, yup. Still have the Kernighan & Ritchie C book. My first "Hello World" was on a *printing* terminal off of a Vax, running 300 baud....

Regards,

Ardis
6) Questions and Answers : Windows : Installing BOINC as a service (Message 19670)
Posted 2 Jul 2006 by Ardis
Post:

did you stop the service before attempting to install 5.4.9? If the boinc manager is running you'll have trouble. Try exiting the manager with "File", "exit" (not the red X). Then try to install it.


Tony,

Thanks for your comments. Pretty sure 5.2.13 was uninstalled before I tried to install 5.4.9, but I just tried it again to make sure. Same result, error 1920, not enough permissions to install as a service. That sounds like a Windows restriction, btw, not BOINC. Didn't have any trouble installing 5.2.13 as a service, same OS (with a few updates). Currently on a single user install, which is running fine. As a service, though, BOINC will restart itself after a power outage, which happens this time of year, so I wouldn't have to stay home when I think the power might go out. 8^)

Regards,

Ardis
7) Questions and Answers : Windows : Installing BOINC as a service (Message 19592)
Posted 30 Jun 2006 by Ardis
Post:
Just read today on another board that the BOINC client doesn't update itself, so I got 5.4.9 and tried to install it as a service, same as 5.2.13 was. That attempt produced an error message saying I didn't have sufficient permissions to install a service, so I rebooted into safe mode, logged in as admin, and tried again. Another error, can't install a service in safe mode. The only way I know how to get to admin is through safe mode, so I'm stuck. Suggestions?

Thanks for your consideration,

Ardis

P.S. There must be a good reason that BOINC doesn't update itself, but oh, the shame, I was more than a month out of date, and didn't even know...






©2021 University of Washington
https://www.bakerlab.org