I have Mcafee Virusscan and was on Boinc 4.09. Worked fine. I downloaded and updated to Boinc 4.11 and later tried 4.10 and on each occasion, 4.10 and 4.11 reacted badly with Mcafee under XP and 98SE so that the machines locked up the processor was so pushed.

I suggest that you dont go past 4.09 if you have Mcafee until Boinc people fix this one.

gregh

Are you a Beta or Alpha tester?

If not you shouldn't be running those versions untill new Boinc is released
Reinstall 4.09

Timmy

---

> gregh
>
> Are you a Beta or Alpha tester?
>
> If not you shouldn't be running those versions untill new Boinc is released
> Reinstall 4.09
>

Easy to say. It was easier to go UP than down at one stage.

> gregh
>
> Are you a Beta or Alpha tester?
>
> If not you shouldn't be running those versions untill new Boinc is released
> Reinstall 4.09
>
> Timmy

Everyone is allowed to DL the Alpha executables as long as the understanding is that there is higher risk, and a strictly factual bug report (like this one) is welcome, but whining, calling the developers names etc. is not welcome at all.

A few more details about what McAffee is saying would be appreciated. Is it reporting a virus of some sort? If so, which one? If so, which application? Is McAffee trapping? Is BOINC trapping? Is S@H trapping?

> > gregh
> >
> > Are you a Beta or Alpha tester?
> >
> > If not you shouldn't be running those versions untill new Boinc is
> released
> > Reinstall 4.09
> >
> > Timmy
>
> Everyone is allowed to DL the Alpha executables as long as the understanding
> is that there is higher risk, and a strictly factual bug report (like this
> one) is welcome,

....which it is, as you pointed out.

> but whining, calling the developers names etc. is not welcome
> at all.

....which no-one did in this thread, up to and including this reply. I dont see a problem with my original post saying "until they fix it".

>
> A few more details about what McAffee is saying would be appreciated. Is it
> reporting a virus of some sort? If so, which one? If so, which application?

It isnt actually REPORTING a virus but it is setting McShield off as if it is continually scanning an application. If I disable Mcafee, then the CPU usage goes down to acceptable once more. Obviously that is not something anyone should do. When I had both going, things would wind up so far as CPU use is concerned, to the point where it would be unable to be used and a reset button job was all you could do to fix it. Interestingly it took minutes to get that far. Further, a P2 400Mhz I have with Mcafee and Boinc using Win98SE didnt suffer anywhere near as badly as a P3 600Mhz with the same on it. The P3 went just as dead as my P3 1.13Ghz laptop with Win XPSP2 and my desktop P4 1.7Ghz also with Win XPSP2. No idea why the P2 would have some processing grunt left and the others wouldnt.

> Is McAffee trapping? Is BOINC trapping? Is S@H trapping?

BOINC and Mcafee are the ones competing. Ending the BOINC process on XP just jumped S&H 4.05 to the top of the queue then and it and Mcafee competed. None of this happens on any machine with Boinc 4.09 and Seti 4.05.

BTW, just a request aside from this bug report - could someone start thinking about jumping the BOINC revision numbers up to, say 20.x for example? It is starting to get a little confusing with Seti at 4.05 and Boinc at 4.09. Thanks.

Greg.

> BOINC and Mcafee are the ones competing. Ending the BOINC process on XP just
> jumped S&H 4.05 to the top of the queue then and it and Mcafee competed.
> None of this happens on any machine with Boinc 4.09 and Seti 4.05.

By killing the GUI and stating the problem still exists with only the project application running, you've just ruled out the BOINC CC being the issue. (you say it left the process list?)

This would seem to point to the new 4.05 project application. But you also state that you have no problems running 4.09 with 4.05. A bit of a stumper.

Just for kicks, try killing S@H and see what happens. (this might void the WU, but it's all in the name of science right?) Does the problem still exist with only the 4.10 or .11 GUI and mcafee running?

Also, has mcafee done any updates recently? (other than virus def updates)

>BTW, just a request aside from this bug report - could someone start thinking
>about jumping the BOINC revision numbers up to, say 20.x for example? It is
>starting to get a little confusing with Seti at 4.05 and Boinc at 4.09. Thanks

Unfortunately all applications must have the same major version as BOINC, 4.xx. The BOINC client will not accept applications that do not match.

John Keck -- BOINCing since 2002/12/08 --

What version of McAfee are we talking about here? I have McAfee v7 Enterprise running on Win2k, Win2k3, and XP Pro. All work properly.

I suppose I could try McAfee v4.51 and see if I can reproduce the problem.

> > BOINC and Mcafee are the ones competing. Ending the BOINC process on XP
> just
> > jumped S&H 4.05 to the top of the queue then and it and Mcafee
> competed.
> > None of this happens on any machine with Boinc 4.09 and Seti 4.05.
>
> By killing the GUI and stating the problem still exists with only the project
> application running, you've just ruled out the BOINC CC being the issue. (you
> say it left the process list?)

Wrong. I did no such thing. If what you contend were the case, then it would be the same with every single BOINC. As stated before, Boinc 4.09 doesnt have this problem even WITH the latest S@H installed.

>
> This would seem to point to the new 4.05 project application. But you also
> state that you have no problems running 4.09 with 4.05. A bit of a stumper.
>
> Just for kicks, try killing S@H and see what happens. (this might void the WU,
> but it's all in the name of science right?) Does the problem still exist with
> only the 4.10 or .11 GUI and mcafee running?
>
> Also, has mcafee done any updates recently? (other than virus def updates)
>

No updates from Mcafee are the issue here. Boinc 4.09 still running happily right now with everything else.

> >BTW, just a request aside from this bug report - could someone start
> thinking
> >about jumping the BOINC revision numbers up to, say 20.x for example? It
> is
> >starting to get a little confusing with Seti at 4.05 and Boinc at 4.09.
> Thanks
>
> Unfortunately all applications must have the same major version as BOINC,
> 4.xx. The BOINC client will not accept applications that do not match.

Surely THAT bit of coding would be the EASIEST thing to change in the whole program?

> What version of McAfee are we talking about here? I have McAfee v7 Enterprise
> running on Win2k, Win2k3, and XP Pro. All work properly.
>
> I suppose I could try McAfee v4.51 and see if I can reproduce the problem.

4.51 hasnt been supported for ages now!

I am talking about Mcafee 9.010 with engine 4.3.20 and DAT 4.0.4396. It is the latest there is.

> > >BTW, just a request aside from this bug report - could someone start
> > thinking
> > >about jumping the BOINC revision numbers up to, say 20.x for example?
> It
> > is
> > >starting to get a little confusing with Seti at 4.05 and Boinc at
> 4.09.
> > Thanks
> >
> > Unfortunately all applications must have the same major version as
> BOINC,
> > 4.xx. The BOINC client will not accept applications that do not match.
>
> Surely THAT bit of coding would be the EASIEST thing to change in the whole
> program?
>
>
Not really. The idea for having major version changes is that they are not compatible with previous versions in some fundamental way. In the case of 3 to 4 the version 4 programs send a heartbeat signal back and forth from the manager to the worker. Version 3 (and lower) did not do this so a version 4 manager trying to run a version 3 application would not know the application was running and try to start it again. A version 4 application run by a version 3 manager would not know the manager was running and exit automatically.

If I remember correctly 1 to 2 and 2 to 3 were file format changes. 0 to 1 was a test to see if the manager would properly reset all projects and refuse to work with the 0 version applications.

John Keck -- BOINCing since 2002/12/08 --

> > >BTW, just a request aside from this bug report - could someone start
> > thinking
> > >about jumping the BOINC revision numbers up to, say 20.x for example?
> It
> > is
> > >starting to get a little confusing with Seti at 4.05 and Boinc at
> 4.09.
> > Thanks
> >
> > Unfortunately all applications must have the same major version as
> BOINC,
> > 4.xx. The BOINC client will not accept applications that do not match.
>
> Surely THAT bit of coding would be the EASIEST thing to change in the whole
> program?
>
Umm, No. The reason for the major version change is a change in the communications protocols that is NOT backwards compatible.

> Wrong. I did no such thing. If what you contend were the case, then it would
> be the same with every single BOINC. As stated before, Boinc 4.09 doesnt have
> this problem even WITH the latest S@H installed.

OK then, lets slow down a bit...

Did you not say that you terminated the BOINC GUI process and left the S@H application running, then S@H and mcafee started competing?

"BOINC and Mcafee are the ones competing. Ending the BOINC process on XP just jumped S&H 4.05 to the top of the queue then and it and Mcafee competed. None of this happens on any machine with Boinc 4.09 and Seti 4.05."

Is this in fact not the case?

If the BOINC GUI is now gone from the process list, it is no longer running. If now S@H is competing with mcafee, this infers the problem is actually the WU interfering in the first place, not the GUI. As terminating the GUI should have alleviated the problem if it were the cause.

Still with me here?

So why would 4.05 cause a problem under 4.10-.11 that doesn't occur with 4.09? That's the big question. Maybe there is something new in the interactions between GUI and project app. Perhaps the way the project app is called by 4.10-.11, or there is some new communication method that puts the project app in a state that triggers mcafee. This would mean 4.09 and 4.05 could be fine, yet 4.10-.11 and 4.05 might not be.

So why not try what I suggested and try killing the WU? Does the problem still exist with the 4.10-.11 GUI alone running? This would help dev pinpoint where the actual problem lies.