BOINC seemsto be terminiting runs,claiming that they are complete, which is impossible. Recently I have switched from Norton to Kaspersky. Could this be the problem ?

Regards, George Macarthur

It looks like your last completion was from Feb 11th. Everything downloaded or running after that died at some point with lots of \"no heartbeat from core client\" messages. When did you switch AV programs? Did you make any changes to your hardware or software sometime around mid February? Changes to your BOINC configuration?

It looks like your last completion was from Feb 11th. Everything downloaded or running after that died at some point with lots of \"no heartbeat from core client\" messages. When did you switch AV programs? Did you make any changes to your hardware or software sometime around mid February? Changes to your BOINC configuration?

I did switch from Norton to Kaspersky on Feb 19. No deliberate changes were made to the BOINC configuration. If this is indeed the problem, what should I do?. I will suspend operations in the meantime.

Regards, George Macarthur

Within Kapersky, is there a way to exclude the BOINC and BOINCData folders from all scanning (on-demand and scheduled)? I would try that first. A number of users have done this with various AV programs to decrease the chances of screwy updates from the AV vendor deciding BOINC or some science application is malware. Of course it\'s up to you if you feel comfortable doing that.

Within Kapersky, is there a way to exclude the BOINC and BOINCData folders from all scanning (on-demand and scheduled)? I would try that first. A number of users have done this with various AV programs to decrease the chances of screwy updates from the AV vendor deciding BOINC or some science application is malware. Of course it\'s up to you if you feel comfortable doing that.

A computer maintenance engineer has given all the BOINC files and Hadley files \"trusted\" status in terms of Kaspersky but the runs are still being terminated, typically after 12-24 hours. He has asked me to ask you which ports BOINC uses. Though he is very experienced, he is mystified as to what is happening.

Regards, George Macarthur

A computer maintenance engineer has given all the BOINC files and Hadley files \"trusted\" status in terms of Kaspersky but the runs are still being terminated, typically after 12-24 hours. He has asked me to ask you which ports BOINC uses. Though he is very experienced, he is mystified as to what is happening.

Regards, George Macarthur

For external communication with projects over the internet, BOINC uses HTTP port 80. Very rarely, it also uses HTTPS port 443, but so far as I know only WCG uses that one.

In addition, the BOINC Manager and the BOINC Core Client (the \'daemon\') communicate internally on localhost port 31416: this post is also used on a local trusted network if, for example, you wish to view/control a daemon on one machine from a manager on a different machine.

All port usage is TCP/IP only.

Here are a couple of suggestions for you to try George.

In the BOINC data directory there\'s a file stderrdae.txt. If the date modified on the file is close to your most recent failure the file is likely to contain something relating to the underlying problem. If that\'s the case extract the relevant messages and post them here if you need help interpreting them.

If that file reveals nothing try enabling debugging of the BOINC client to decode the messages passed to/from the running application. Create the file cc_config.xml in your BOINC data directory, open it with notepad and paste the 6 lines I\'ve quoted below:

<cc_config>
<log_flags>
<app_msg_receive>1</app_msg_receive>
<app_msg_send>1</app_msg_send>
</log_flags>
</cc_config>

In BOINC manager\'s advanced view select the menu option Advanced - Read config file and you should start seeing messages starting with [app_msg_send] and [app_msg_receive]. As CPDN is your only BOINC project and your system has a single CPU restricted to one task per day it should be fairly easy to identify what happens when the next heartbeat failure happens; copy the last few debug messages and post them here.

To disable the debug messages change the values in cc_config.xml to 0 and reload the file.

---

"The ultimate test of a moral society is the kind of world that it leaves to its children." - Dietrich Bonhoeffer

A computer maintenance engineer has given all the BOINC files and Hadley files \"trusted\" status in terms of Kaspersky but the runs are still being terminated, typically after 12-24 hours. He has asked me to ask you which ports BOINC uses. Though he is very experienced, he is mystified as to what is happening.

Regards, George Macarthur

For external communication with projects over the internet, BOINC uses HTTP port 80. Very rarely, it also uses HTTPS port 443, but so far as I know only WCG uses that one.

In addition, the BOINC Manager and the BOINC Core Client (the \'daemon\') communicate internally on localhost port 31416: this post is also used on a local trusted network if, for example, you wish to view/control a daemon on one machine from a manager on a different machine.

All port usage is TCP/IP only.