| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: Africa Rainfall Project
Thread: For Current Volunteers: Advance Information on Our Newest Project |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 164
|
|
| Author |
|
|
Mike.Gibson
Ace Cruncher England Joined: Aug 23, 2007 Post Count: 12419 Status: Offline Project Badges: 
|
An extra thought.
Are you getting updates loaded automatically overnight? That could cause your machine to reboot and lose your crunching. Change your update settings to download automatically but leave you to reboot when you wish that is when an ARP has just checkpointed. Mike |
||
|
|
leloft
Cruncher Joined: Jun 8, 2017 Post Count: 23 Status: Offline Project Badges: 
|
Please could you clarify a small but important technical point about checkpoints. The two machines running ARP are configured with 'swappines=0'. Is there a way I can create a 'boinc-swap' in my 50G build partition so that I can force a checkpoint by 'suspending' before a reboot and thus avoid losing data?
Thanks. |
||
|
|
Acibant
Advanced Cruncher USA Joined: Apr 15, 2020 Post Count: 126 Status: Offline Project Badges: 
|
Is there a way I can create a 'boinc-swap' in my 50G build partition so that I can force a checkpoint by 'suspending' before a reboot and thus avoid losing data? If you were just hibernating that would write RAM to disk. Rebooting is going to be trickier. If this is a frequent occurrence and you really want to work on ARP without losing data due to only having checkpoints every 12.5%, it would be good to look at running BOINC inside a virtual machine. Then you can hibernate that VM and its memory state isn't going to be wiped out by the host rebooting. If you go this route, make sure virtualization is enabled in your BIOS to keep the performance impact to the absolute minimum.Thanks.  |
||
|
|
leloft
Cruncher Joined: Jun 8, 2017 Post Count: 23 Status: Offline Project Badges:
|
If this is a frequent occurrence and you really want to work on ARP without losing data due to only having checkpoints every 12.5%, Thanks for your reply. Thee two machines are only rebooted when necessary, approx. every couple of months, so that's not a problem. It's largely theoretical, but I'm running 24 cores so that's a potential equivalent of 3 whole ARP tasks lost. Be good to force a checkpoint, write it all to disk and pick it up after a reboot. |
||
|
|
|