Error Message "No Heartbeat on Core Client"
16 Apr 2008 5:39:24 UTC
Topic 193631
(moderation:
I keep noticing that the message log shows a dll initalization error and the client restarting. I just completed a wu on this system and a look at the wu gives me this message:
175, 176, 177, No heartbeat from core client for 30 sec - exiting
2008-04-15 16:04:36.9218 [normal]: Built at: Feb 25 2008 17:11:17
2008-04-15 16:04:36.9218 [normal]: Start of BOINC application 'projects/einstein.phys.uwm.edu/einstein_S5R3_4.36_windows_intelx86.exe'.
2008-04-15 16:04:36.9375 [debug]: Set up communication with graphics process.
2008-04-15 16:04:37.4375 [debug]: Reading SFTs and setting up stacks ... done
2008-04-15 16:05:19.3593 [debug]: Successfully read checkpoint
2008-04-15 16:05:19.3750 [debug]: Total skypoints = 1202. Progress: 175,
$Revision: 1.119 $ REV:$Revision, OPT:3, SCVAR:9, SCTRIM:2, HOTVAR:3, HOTDIV:0, HGHPRE:7, HGHBAT:2
I noticed that while using the performance meter in XP that einstein stopped. Looking at the task manager it was no longer active and only boinc was. I've looked every where I can think of to try and track down whatever this dll is and can't find any information as to what one is the problem. Much less what is causing einstein to shut down.
Any help would be appreciated.
PS... I noticed this when I was running seti and the work always finished without a problem. Apparently it also affects einstein.
Error Message "No Heartbeat on Core Client"
)
Do you happen to use Internet time servers to automatically adjust your PC time? If so, just ignore the messages. It's not a project specific problem, it's caused by boinc's heartbeat mechanism to check if the application is still working. If it doesn't respond for longer than 30 seconds, boinc assumes that it's "dead" and drops the process. Unfortunately, that also happens when the PC clock is adjusted in the "wrong" direction. Just yesterday I saw this happen to einstein and seti simultaneously.
With older boinc versions, the message used to be something like "... exited with zero status but no 'finished' file... you might need to reset the project". Now you are told that a reboot is sufficient and that there had been a DLL initialization error. In my case that message occurred exactly when the PC clock was adjusted, so I ignored it.
There have been several threads on this forum and the boinc message boards, but I can't seem to find them now (lack of keywords :-).
Gruß,
Gundolf
Computer sind nicht alles im Leben. (Kleiner Scherz)
RE: With older boinc
)
Yes, I believe that warning message has been replaced with the DLL initialization warning and both are harmless.
This is the most recent occasion where I wrote about this warning message.
Cheers,
Gary.
I did a search too before I
)
I did a search too before I posted and couldn't get any results. I have 4 systems running here and it is only happening on the slowest one (XP2600 Throughbred core). Its runing XP Pro, basically sits on the shelf and ignores everyone and has 512 megs of ram. I noticed that I was also getting error messages in the Event viewer and it turned out to be something related to McAfee's updater. I tracked that down, called McAfee and we did some tests and everything was okay. 30 minutes later the error message came back. I uninstalled McAfee and no more errors there. But still getting them in Boinc. If the dll init. statement is misleading then I'm going to quit trying to figure out what the heck is going on.
I am running the latest client and it happens whether or not I am using the standard or optimized client. Don't know if this is related or not but I have onboard video. Maybe I'll try sticking a graphics card in and seeing what that does.
RE: I did a search too
)
The DLL message is relatively recent and may not be mentioned on these boards. I would expect there would be more on the BOINC boards about it. I believe I read somewhere that the "zero exit status" message was thought to be too scary so was turned into the DLL message in a recent BOINC version. 5.10.28 still gives the old version of the message. I believe it's the same thing and it's only a benign warning.
I see the message on Athlon XPs as well. I've just checked a machine known to display the warning. There were three occurrences in the last 24 hours in its messages log. The machine has been crunching 24/7 for the last three years and has accumulated close to 400K credits in that time with no problems that I'm aware of.
I don't think it's supposed to be misleading - Rom Walton probably knows exactly what it means :-), and it may well be documented over on the BOINC boards. I don't have time to visit there often so I don't know. No doubt Jord may know so he might be able to give a pointer. I think the previous message may have been deemed to be misleading because of the reference to resetting the project, which was one reason why it was changed.
I would say unrelated. My machine has no onboard video but still has the warning message at regular intervals.
Cheers,
Gary.