I recently set my connect to server preference to 4 days with the understanding that the boinc client would connect to the work unit server and download enough units to keep my computer working for about 4 days. This doesn't seem to work--even though I've restarted the client several times. I currently have a single work unit waiting for each processor. Any ideas? Is this a glitch in the boinc client or is something else going on?

Using 5.4.9 on a G5 PowerMac. Thx.

I recently set my connect to server preference to 4 days with the understanding that the boinc client would connect to the work unit server and download enough units to keep my computer working for about 4 days. This doesn't seem to work--even though I've restarted the client several times. I currently have a single work unit waiting for each processor. Any ideas? Is this a glitch in the boinc client or is something else going on?

Have you done a update project within the boinc manager ? - Please, select the projct and click on the "update projects" button, to let your boinc client download the new preferences you have made on the project web site.

---

I recently set my connect to server preference to 4 days with the understanding that the boinc client would connect to the work unit server and download enough units to keep my computer working for about 4 days. This doesn't seem to work--even though I've restarted the client several times. I currently have a single work unit waiting for each processor. Any ideas? Is this a glitch in the boinc client or is something else going on?

There’s apparently a new bug in the website configuration (they’ve been tinkering with it quite a bit recently) that’s preventing proper updating of preferences. According to the Technical News of Oct. 12, it’s been squashed but the repaired code hasn’t been implemented yet. (They might do that during tomorrow’s scheduled outage.)

Possible solutions include: just waiting a few days; signing up with another project (a BOINC account’s General Preferences can be updated at any project website); and ‘hacking’ your local XML files, to change the setting manually (I can’t advise on exactly how to do this, but someone else here probably can).

---

The cache is working for holding a fresh work unit for each processor--so no problems unless there is a server outage which stops the feed. I did "update projects" to no effect. Perhaps pref settings are not affecting the server and that an update there will fix things.

I was assigned several tasks to complete. Once completed, I have not received any additional work to be run on my computer. I keep getting a message stating that communication is being deferred and then there is a time unit that counts backwards to 0. When I get to zero nothing happens. I have tried everything but am getting no where.

There’s apparently a new bug in the website configuration (they’ve been tinkering with it quite a bit recently) that’s preventing proper updating of preferences. According to the Technical News of Oct. 12, it’s been squashed but the repaired code hasn’t been implemented yet. (They might do that during tomorrow’s scheduled outage.)

Odysseus, do you know any further ? - Was the problem fixed by at the last maintainance ?

---

Odysseus, do you know any further ? - Was the problem fixed by at the last maintainance ?

Sorry, I have no idea; all I have to go on is what’s been said in these forums and on the Technical News page. I haven’t changed any preferences here for weeks, maybe even months.

---

As of this morning, prefs for setting number of days of work units (connect to server) still isn't working.

---

Hello,

I have the same problem of cache on a Mac Pro. And 5.4.9 ...

---

Hello,

I have the same problem of cache on a Mac Pro. And 5.4.9 ...

Since the maintainance two weeks ago it works agian. - Have you updated the project via the boinc manager ?

---

Since the maintainance two weeks ago it works agian. - Have you updated the project via the boinc manager ?

I've updated the project via BOINC manager a bunch of times (and restarted BOINC, reset prefs) and it is still not working.

Here is a list of the prefs that are not working on the Mac OSX BOINC client version (5.4.9) (the ones that I've tried to set).

1) Do work while computer is running on batteries?

That is set to NO on my laptop, it is not respected.

2) Do work while computer is in use?

This is set to NO on all my computers. It is not respected--computers keep chugging away at 100%.

3) Do work only after computer is idle for 3 minutes

This isn't respected--as mentioned in 2, computer at full blast all the time.

4) On multiprocessors, use at most 1 processors

This doesn't seem to make any difference.

5) Connect to network about every
(determines size of work cache; maximum 10 days) 4 days

This is the PREF that this thread is about. It doesn't make any difference--still only have 1 or 2 units waiting no matter if I have connect every hour or ever 4 days or whatever.



So, it looks like there are a bunch of prefs that really make no difference on laptops or desktops. I'm running BOINC on G5s and Intel Macs, all running 10.4.8 with no difference in the way prefs make the app behave.

It would be nice to have these prefs working. Best wishes,

Nathan

---

Nathan, you could do a workaround and change your prefs localy :
- stop the boinc client
- copy the file global_prefs.xml in the directory /Library/Application Support/BOINC Data to the file global_prefs_override.xml
- change the preferences in the file global_prefs_override.xml
- start the boinc client

Btw. it did not harm your normal work, if you let the boinc client running while you are working. The boinc client starts the SETI application with the lowest priority, so if your applications need CPU, the OS reduces the CPU from the SETU app and gives the full CPU to your application.
My expiriance is, that if you has enought memory, the applications feel fast as without SETI runnig.

---

Hello,

I have the same problem of cache on a Mac Pro. And 5.4.9 ...

Since the maintainance two weeks ago it works agian. - Have you updated the project via the boinc manager ?

Yes I did....

It's a new computer, it works since 4 days .

I have a PowerMac G5 with an older BOINC 5.4.9 : no problem ...

EDIT: Oups updating the project does not work on the powerMac .


EDIT2: The Mac Pro does not have a file global_prefs_override.xml ???

---

The Mac Pro does not have a file global_prefs_override.xml ???

There isn’t one by default, but you can make your own—it’s easiest to start by copying global_prefs.xml as Dotsch suggested.

---

No file global_prefs.xml on this computer...

---

I tried to make a file global_prefs.xml by copiying the global_prefs.xml from another computer, BOINC wanted more than 2 000 000 seconds of work.
But the answer from the server was: "request incomplete".
I make (made?..) a file global_prefs_override.xml... Did not work.."request incomplete"...
So I removed the global_prefs.xml. And it works. BOINC is downloading many and many wu.
Now it seems to be OK .

---

Well this still seems to be a problem. I'm running on a Mac Pro and an iMac G5 as well as a couple of Windows laptops and every Tuesday for about 8 hours I can't connect.

This is a real pain now because the Pro finishes a WU for each core every 2 hours. I get one WU cached for each core but I'm requesting enough for 10 days. I've tried 2 days also, makes no difference.

Any ideas? Cheers

Dave

Well this still seems to be a problem. I'm running on a Mac Pro and an iMac G5 as well as a couple of Windows laptops and every Tuesday for about 8 hours I can't connect.

This is a real pain now because the Pro finishes a WU for each core every 2 hours. I get one WU cached for each core but I'm requesting enough for 10 days. I've tried 2 days also, makes no difference.

Paradoxically enough, a ten-day cache can be a problem. A recent version of the BOINC scheduler will kick into ‘panic mode’ as soon as it sees work that’s due before the next scheduled connection. Considering that this project’s deadlines can be as short as 3.5 d (? something like that, anyway), the likelihood of this happening becomes quite high with CIs longer than that. I would recommend reducing it to no more than two or three days, and leaving it there for a while. What version of BOINC are you running? What do the hosts’ efficiency statistics look like? Do you run any other projects?

Unfortunately we may be in for a prolonged outage: apparently thumper has had a serious hardware failure—see the technical news forum. Matt seems optimistic about a replacement being forthcoming, but you never know …

Well I've tried it on lower settings and never saw more than one WU queued.

I guess further experimentation now will have to wait a little!

Well I've tried it on lower settings and never saw more than one WU queued.

I guess further experimentation now will have to wait a little!

If you end up downloading any of the short (4 day) WUs, a connect setting of anything over 1 will cause BOINC to enter panic mode. That causes BOINC to stop downloading any more WUs until all the short WUs are finished.

Bottom line: So long as S@H continues sending out WUs with a 4 day due date, a connect setting over 1 day will cause a smaller work queue, not larger. Yes...it's the opposite of what makes sense. But there it is.

BOINC client version 5.9 has a new setting in it allowing to manually add an additional number of days worth of work to your queue. But be warned, it may put you in a position where can miss deadlines.

---

Dublin, California
Team: SETI.USA