As usual, not for the faint of heart.

Downloading as we speak.

Warning: 4.42 is developmental software. Stick with the official 4.19 or 4.25 versions if you're not willing to take the risks and report the bugs.

thanks
tony

Heck why not, I'll give it a go... have been upgrading so many times this past couple of weeks whats one more... btw, I found 4.40 fairly stable.
My biggest problem at the moment is getting my CP stats listed along side my SETI & Einstein.. but thats another 'long' story :(

---

Cheers

Gav



Nothing is 'fool-proof', someone will always invent a better fool!

I'm currently downloading 4.42, and if:

1. There are bugs, I'm reporting them and going back to 4.25
2. There aren't bugs, I'm keeping it

---

<a href="http://tinyurl.com/9hemz"><img src="http://www.boincsynergy.com/images/stats/comb-1441.jpg"></img></a>

AMAZING! It appears this one has no bugs on my system...


I was starting to lose faith that you guys could produce a new version w/ no bugs (I know I konw, all dev versions use at your own risk, bla bla, but 4.42 is a shining example of functionality for me).

---

<a href="http://tinyurl.com/9hemz"><img src="http://www.boincsynergy.com/images/stats/comb-1441.jpg"></img></a>

Wow, NO connection failed message, project update works. Give her a shot

tony

[edit] Hmm don't remember getting the CLI access to the firewall, restarting puter. Don't see the the version change in the message log. Hang on.


[edit again] I take back some of what I said. after reboot I got the "Connection Failed" message. Then only after I clicked "project update" did the firewall ask for permission to allow access for the CLI.

Project update does still work. I see the version change in the message log.

I'm getting tired of having to enter "localhost" every single time. can we fix that next?

Say Tony, have you thought about it that you have to fill in "localhost" every time because of your firewall? What are you using? Firewall wise then, not asking any personal questions. ;)

---

Wow, NO connection failed message, project update works. Give her a shot

tony

[edit] Hmm don't remember getting the CLI access to the firewall, restarting puter. Don't see the the version change in the message log. Hang on.


[edit again] I take back some of what I said. after reboot I got the "Connection Failed" message. Then only after I clicked "project update" did the firewall ask for permission to allow access for the CLI.

Project update does still work. I see the version change in the message log.

I'm getting tired of having to enter "localhost" every single time. can we fix that next?

1. I dont find I need to actually enter 'localhost' - just click Ok with
blank fields works for me.

2. What firewall? I have ZoneAlarm. I can right-click the program and select
'Changes Frequently' - stops it version checking, just uses tthe program name.

---

Say Tony, have you thought about it that you have to fill in "localhost" every time because of your firewall? What are you using? Firewall wise then, not asking any personal questions. ;)

Nope hadn't thought of it. I only did it on rare occasions from 4.20 to 4.32, then with 4.35 to 4.41 it's pretty much all the time. Fire wall hasn't changed. What has changed is the timing of when i'm prompted to give the CLI permission. On the 4.20 to 4.32 it was instanly after the prompt to give the permission to the Boinc Manager. On these, there's a Lag, and the Connection failed occurs directly after the Boinc Manager prompt. The Prompt for the CLI permission comes much later (when an actual attempt to communicate is made). I think some change in the programming is delaying the CLI from trying to access my network for a time period which exceeds the Error message time.

thanks for thinking of me though.

tony

@ ageless, I'm using Mcafee internet suite for 2005. build 6.1.6184

Wow, NO connection failed message, project update works. Give her a shot

tony

[edit] Hmm don't remember getting the CLI access to the firewall, restarting puter. Don't see the the version change in the message log. Hang on.


[edit again] I take back some of what I said. after reboot I got the "Connection Failed" message. Then only after I clicked "project update" did the firewall ask for permission to allow access for the CLI.

Project update does still work. I see the version change in the message log.

I'm getting tired of having to enter "localhost" every single time. can we fix that next?

1. I dont find I need to actually enter 'localhost' - just click Ok with
blank fields works for me.

2. What firewall? I have ZoneAlarm. I can right-click the program and select
'Changes Frequently' - stops it version checking, just uses tthe program name.

I wish I could do the changes frequently thing, but all I do is click 'Allow Access' under McAfee. And I've never gotten the connection failed thing (actually, I have, but it was because I took too long to respond to the firewall, so I would respond and then restart BOINC). Tony, I think the problem is in your firewall too....

---

<a href="http://tinyurl.com/9hemz"><img src="http://www.boincsynergy.com/images/stats/comb-1441.jpg"></img></a>

Tony, I think the problem is in your firewall too....

I think I'd give that thought more time, If this wasn't such a widespread issue amongst users. If there was only a couple of us, or it happened to only people with my version of Mcaffe, then I'd seriously consider it. The change in event timing seems more likely. But hey, I'm willing to listen.

thanks

tony

Tony, I think the problem is in your firewall too....

I think I'd give that thought more time, If this wasn't such a widespread issue amongst users. If there was only a couple of us, or it happened to only people with my version of Mcaffe, then I'd seriously consider it. The change in event timing seems more likely. But hey, I'm willing to listen.

thanks

tony

I don't pretend to be the know-all person on computers, so it may very well be not. I'm not quite aware how widespread the problem is. Anyways...good luck with it.

---

<a href="http://tinyurl.com/9hemz"><img src="http://www.boincsynergy.com/images/stats/comb-1441.jpg"></img></a>

I had a problem earlier today with wu's from 4.30 disappearing after they had uploaded. Earlier today I upgraded to 4.42 and finished another PPWU (1783186). After it finished crunching it uploaded and started on the next one. I hit manual update and see that it contacted the server. But the WU has disappeared...

5/16/2005 9:51:20 PM|ProteinPredictorAtHome|Computation for result h0008A_1_43251_3 finished
5/16/2005 9:51:20 PM||schedule_cpus: must schedule
5/16/2005 9:51:20 PM|ProteinPredictorAtHome|Starting result h0008A_1_43186_1 using mfoldB125 version 4.28
5/16/2005 9:51:22 PM|ProteinPredictorAtHome|Started upload of h0008A_1_43251_3_0
5/16/2005 9:51:22 PM|ProteinPredictorAtHome|Started upload of h0008A_1_43251_3_1
5/16/2005 9:51:23 PM|ProteinPredictorAtHome|Finished upload of h0008A_1_43251_3_0
5/16/2005 9:51:23 PM|ProteinPredictorAtHome|Throughput 10954 bytes/sec
5/16/2005 9:51:23 PM|ProteinPredictorAtHome|Started upload of h0008A_1_43251_3_2
5/16/2005 9:51:24 PM|ProteinPredictorAtHome|Finished upload of h0008A_1_43251_3_1
5/16/2005 9:51:24 PM|ProteinPredictorAtHome|Throughput 27959 bytes/sec
5/16/2005 9:51:25 PM|ProteinPredictorAtHome|Finished upload of h0008A_1_43251_3_2
5/16/2005 9:51:25 PM|ProteinPredictorAtHome|Throughput 63446 bytes/sec
5/16/2005 9:51:47 PM||request_reschedule_cpus: project op
5/16/2005 9:51:47 PM||schedule_cpus: must schedule
5/16/2005 9:51:48 PM|ProteinPredictorAtHome|Sending scheduler request to http://predictor.scripps.edu/predictor_cgi/cgi
5/16/2005 9:51:49 PM|ProteinPredictorAtHome|Scheduler request to http://predictor.scripps.edu/predictor_cgi/cgi succeeded


It seems to happen no matter what version I am using

I'm headed to work so I'll leave it crunching CPDN until the morning. I don't want to lose any more wu's.

It seems to happen no matter what version I am using

I guess I don't understand the problem. The log you posted looks completely normal to me. It finished a wu, then uploaded it, then reported it back. this should remove the WU from the "Work" tab. Maybe I just don't get it. Have you checked the results page from that WU to see if they got it?

tony

Anyone else seeing 4.42 still being reluctant to connect to the schedulers? Seems quite a coincidence that all the projects are having no schedulers responding, then switch back to 4.35 and the schedulers are responding again...

---

Went from 4.41 -> 4.42 and a bunch of ready to be reported WU's were automatically reported. Schedulers seem to be responding fine for me heffed.

---

Kolch - Crunching for the BOINC@Australia team since July 2004.
Search for your own intelligence...

Anyone else seeing 4.42 still being reluctant to connect to the schedulers? Seems quite a coincidence that all the projects are having no schedulers responding, then switch back to 4.35 and the schedulers are responding again...

Not yet:

finished seti and it auto uploaded.

5/16/2005 10:14:02 PM|SETI@home|Computation for result 31dc04aa.14083.8752.540910.33_0 finished
5/16/2005 10:14:02 PM||schedule_cpus: must schedule
5/16/2005 10:14:02 PM|SETI@home|Starting result 30dc04ab.18554.27136.384654.130_2 using setiathome version 4.09
5/16/2005 10:14:03 PM|SETI@home|Started upload of 31dc04aa.14083.8752.540910.33_0_0
5/16/2005 10:14:07 PM|SETI@home|Finished upload of 31dc04aa.14083.8752.540910.33_0_0
5/16/2005 10:14:07 PM|SETI@home|Throughput 25384 bytes/sec

then clicked update

5/16/2005 11:50:29 PM||request_reschedule_cpus: project op
5/16/2005 11:50:29 PM||schedule_cpus: must schedule
5/16/2005 11:50:30 PM|SETI@home|Sending scheduler request to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi
5/16/2005 11:50:32 PM|SETI@home|Scheduler request to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi succeeded

seems ok to me so far. I got 6 days worth of work earlier today when I downgraded to 4.32, so I won't be testing the work fetch any time soon on this machine.

tony

I got 6 days worth of work earlier today when I downgraded to 4.32, so I won't be testing the work fetch any time soon on this machine.

I can safely say the work fetch disable function works fine.

---

Cheers

Gav



Nothing is 'fool-proof', someone will always invent a better fool!

Installed 4.42 with no problems on 3 systems.

While _much_ better, it still not perfect when it comes to Update.
When I have 'Ready to Report' work, 7 out 10 clicks of Update result in a
schedule request.The others just cause a CPU schedule event and no
communication attempt. Not meaning 10 consecutive clicks - but single clicks
over a period of hours.

---