Hi Everybody, I have multiple computers running SETI, Einstein, Climate Prediction, etc. Have been doing SETI since 2004. Upgraded Slackware Linux boxes from boinc_5.10.28_pc-linux/gnu to 5.10.45 ... The first time I ran the new version it reset all projects, deleted completed work units that had not been uploaded yet, created new hosts (with the same name as the old host) reset my results to ZERO. I've had to run all over the place and merge hosts, etc. This 'upgrade' experience is just awful. Why don't you TEST a new version before releasing it on unspecting users?

It was tested through BOINC Alpha, with these test results.

If you don't like the end result, Berkeley is always happy to take on new Alpha testers. See how to be an alpha tester for instructions.

By the way, it's always a good idea to upload/report work done before upgrading BOINC.

---

It was tested through BOINC Alpha, with these test results.

If you don't like the end result, Berkeley is always happy to take on new Alpha testers. See how to be an alpha tester for instructions.

By the way, it's always a good idea to upload/report work done before upgrading BOINC.

Thanks for a speedy hard-hitting reply. This is good. I just now had another bad experience with boinc 5.10.45-linux/gnu. I'm running cornwall.sra.net which is a Thinkpad T30 dualboot Windows XP and OpenSUSE 10.2 ... have had this computer for about a year. Ran a Climate Prediction model from beginning to end on the Windows XP side. During that time I booted Windows 95% Linux 5% ... Did several BOINC work units on the Linux side. As soon as the Climate Prediction model was finished I detached from that project, booted to the Linux side, attached to Climate Prediction and started processing a new model on the Linux side. I did that because I much rather run any version of Linux in preference to any version of Windows. OK. Had boincsimap and Einstein set to not get any new work units. I started running Climate Prediction, then 'upgraded' to the new Linux version 5.10.45 ... That version deleted my partially completed Climate Prediction model, downloaded a new Einstein work unit plus a new Climate Prediction model. Now two weeks have passed and it was time to send 'trickle-up' messages to Climate Prediction and the now-completed Einstein unit. The program downloaded two new Einstein work units even though I had that project set to not get any new work units. THIS IS BUGGY! My new Climate Prediction model is about 1.5% done and I don't expect it to complete for close to a year. So, had to detach from Einstein and from boincsimap on this computer. The more I see the new Linux version of Boinc the less I like it. I'm probably too much of a maverick to be a good Alpha tester. I'm still upset because BOINC never produced a version for OS/2. Speaking of which I processed more than 700 Seti work units on the old pre-BOINC seti software. Half were on the OS/2 version half were on Vector Linux and one on Windows 98 (on a Thinkpad 760XD) The one Windows 98 version took a month to process running 24/7 in contrast to the OS/2 version or the Linux version which would take about 20 hours. So, BOINC in its wisdom failed to produce an OS/2 version. Go figure. Thanks again for the reply.

I too have had problems with the new version although I do not think it reset any projects.
When I upgraded to the latest version of 64bit BOINC 5.10.45 the client stops running when I close the manager window whether through the close window menu or the X in the top right I cant' close the window without stopping the client and have leave the window permanently open. Is this a known issue or am I doing something wrong.
I was running 64bit Linux BOINC 5.10.28 originally which had a problem of not closing the client when exiting and had too exit via the advanced menu to stop the client running I under stood this was a known bug.
I am newbie to Linux and still felling my way I am more used to windows!
I am using Ubuntu 7.10 on an AMD CPU

When I upgraded to the latest version of 64bit BOINC 5.10.45 the client stops running when I close the manager window whether through the close window menu or the X in the top right I cant' close the window without stopping the client and have leave the window permanently open. Is this a known issue or am I doing something wrong.

This is normal behaviour. If you exit the BOINC Manager, while you haven't got BOINC installed as a daemon, BOINC will stop. For the 30 minutes (exaggerated) that I ran Mepis Linux I never could put BM into the system tray area. So that's also by default, I think.

A comprehensive How to install BOINC as a daemon (or Sevice) can be found here.

---

I'm still upset because BOINC never produced a version for OS/2.

Since there are only a handful of full-time programmers working on BOINC they can't release it for all operating systems old and new out there. But since it is Open Source, other people have taken it upon themselves to port it to as many other platforms as possible. Including OS/2.

Lars Bausch keeps a list together in this thread with alternative platforms that BOINC is ported to. One of the links in there is:

Yuri Dario ; http://groups.yahoo.com/group/seti-warp/messages/1811?tidx=1
- OS/2 i386

That's mainly a list for Seti & BOINC on OS/2. Other projects may not have OS/2 science applications, especially not those who don't Open Source their application.

---

Thanks
I thought I may be going wrong somewhere but why is there “close window” and “exit” in the “file menu”?
I wish the installing was as easy as Windows. It would be nice if all these tips were referenced too in one place as I did some searching around for advice and missed that.
Actually I have both 32bit and 64 bit installed under different users (I do not run them at the same time ) to compare performance!(about 25% improvement dependent on project)
I run CPDN mainly and had not upgraded the 32 bit.
Dave

Some distros, Ubuntu for sure maybe Fedora too, have BOINC packages that make installing and updating easier.

In Ubuntu BOINC is in the add and remove programs but I know its an older version and it did not say weather it was the 64 or 32 version as CPDN is my main crunching I installed it myself and also due to my ignorance at the time I did not realise it was their.

Hi Srank

Regarding your CPDN models which take so long to complete, it's maddening to lose one when it's well advanced. In the case of the HADCM 160 and 80-year models, all the info you sent to the server in your 10-year zip files will be used by the researchers, so the computing time wasn't all wasted.

I suggest you look at the CPDN README collections which are posted on the CPDN independent forum here near the top. I recommend the section about backing up the contents of the BOINC folder. There's a selection of methods. If you back up regularly and your model crashes or your BOINC installation goes wrong, you restore the backup and continue crunching the same model. Restoring a backup is the only way to rescue a crashed model.

In the README collection about crashes & problems, item #6 by Mike explains how to avoid the usual causes of climate model crashes. And in the collection about running the model you'll find tips on how to run models as fast as possible.

(If you want to post on that forum you'll need to register separately.)

Hope that helps.