Join PrimeGrid
Returning Participants
Community
Leader Boards
Results
Other
drummers-lowrise
|
Message boards :
Generalized Fermat Prime Search :
Terminating because BOINC client heartbeat was lost.
Author |
Message |
|
Hi,
I'm running genefer in a VMWare VM under 64-bit Ubuntu, just for something new to play with, and I'm unable to get any workunit to complete. They are all running when I check at night and not when I get up the next morning. They appear to be failing because " Terminating because BOINC client heartbeat was lost." Can anyone suggest why that would be and what I need to do to prevent it? The BOINC client has never stopped.
Thanks. | |
|
Crun-chi Volunteer tester
 Send message
Joined: 25 Nov 09 Posts: 3247 ID: 50683 Credit: 152,646,050 RAC: 18,212
                         
|
Hi Gary
I am new in Linux world, but have that error many times. I try to find solution on the Internet and it looks like some kind of error in Boinc manager. Every 30 seconds ( or so) Boinc check to see if application is live or not. If there is no response from application you will get that error. But that error is found in many boinc versions, and never was fixed.
I am sorry since I cannot help you more, nut that is what I know about that problem.
You may try with lower CPU freq ( if you do CPU tasks..)
____________
92*10^1585996-1 NEAR-REPDIGIT PRIME :) :) :)
4 * 650^498101-1 CRUS PRIME
2022202116^131072+1 GENERALIZED FERMAT
Proud member of team Aggie The Pew. Go Aggie! | |
|
|
I'm running genefer in a VMWare VM under 64-bit Ubuntu, just for something new to play with, and I'm unable to get any workunit to complete. They are all running when I check at night and not when I get up the next morning. They appear to be failing because " Terminating because BOINC client heartbeat was lost." Can anyone suggest why that would be and what I need to do to prevent it? The BOINC client has never stopped.
The heart beat message was shown a lot of times on your last task. than means that the crunching stopped a lot of times.
I would suggest you to check the boinc preferences. maybe you have it set not to run if cpu use is above some low limit. Or maybe it is a memory issue. The memory comitted to your linux vm is also a bit low (less than 1GB). If boinc can only use a fraction of it, that could maybe cause that error. Try to set both limits higher:
cpu: boinc/tools/preferences/cpu usage tab: while cpu usage is less than "0"
memory: boinc/tools/preferences/disk and memory tab: set memory use in both cases to 80 or 90%.
If you can, try to assign some more memory to the linux vm (half of your host total, for instance), that could also help.
____________
676754^262144+1 is prime | |
|
rroonnaalldd Volunteer developer Volunteer tester
 Send message
Joined: 3 Jul 09 Posts: 1213 ID: 42893 Credit: 34,634,263 RAC: 0
                 
|
Running Genefer in a VMware VM is no problem. I do this on different hosts and the last Genefer unit was done on host dl380-g2.
As you can see, a 64bit linux (DotschUX/1.2) with only 448MB vRAM.
OT:
If you are using VMware Workstation, Player or Fusion (unsure, because never used), using a bunch of memory does not help in any cases. In dependence of the host-OS you will see some strange effects. If you configured the vRAM to high (depended of the product-version between 768-1500MB), your host will hold only a part of vRAM in the pRAM (mostly 300-400MB) and marked the rest as swappable. The host-OS will respond to this immediattely and ave the content of vRAM on the host swap area. Your VM-performance goes down.
Therefore exists the unwritten law for vSMP and vRAM inside a VM: Only as much as needed and not as possible!
For every virtualization product you have to do some homework.
- Disable all energy savings settings. If your host decides to go in sleepmode, he will do this and does not inform the runnings VM(s) about this in all cases. On the other side if a guest/VM decided to go in sleepmode, it can slowdown all other running VMs.
- Set all vDisk (VMware uses the ending VMDK) on the exception list in your antivirus-solution or security-suite. If a file is scanned by your AV-solution this file will be locked for all other processes, your VM can die or you will at least see a bluescreen.
- In awaiting of heavy traffic/usage inside a VM, do not start more VMs than available cores on the host minus one and let 512MB (32bit-host) or 1024MB (64bit-host) free for the host-OS. Your host-OS needs some CPU-cycles and memory for his own work. If there is no free RAM/CPU, the OS will take it independed of all other settings and the VM-performce will go down again.
- Using sparse-disks and/or snapshots. This type of disk is saving a lot of place on the host but fragments in shortest time and you have to shrink them manually. Using snapshots makes the same. All new written content inside a VM will be saved as a VMware sparse-disk on the host-disk and favors fragmentation...
____________
Best wishes. Knowledge is power. by jjwhalen
| |
|
|
Thanks for those suggestions. I thought I'd checked most of those things but I may have missed something. I'll run through them again.
| |
|
|
Still not succeeding, unfortunately. Any other ideas?
| |
|
Message boards :
Generalized Fermat Prime Search :
Terminating because BOINC client heartbeat was lost. |