checkpoint - BEAST 1.10.5

398 views
Skip to first unread message

Luiz JQ

unread,
Oct 29, 2019, 5:47:36 AM10/29/19
to beast-users
Dear colleagues,

I would like to ask you for some help concerning an issue that is probably very trivial (beast version 1.10.5).
As the cluster I use to run beast analyses is time-limited, I want to generate checkpoints to resume the analyses afterward:

beast1 -save_every 25000 -save_state checkpoint347 T437.xml


However, when I try to resume the analysis using the checkpoint file,

beast1 -load_state checkpoint347 T437.xml


I cannot to because beast complains that the log file already exists. The only solution I found is to overwrite or to run the analysis with different names and then to combine the different pieces of the analysis. 

So my question is if there is a way to resume the analysis by appending automatically the new log and tree data to the ones that already exist.

Thanks for the help!

Cheers,

Luiz

Andrew Rambaut

unread,
Oct 29, 2019, 6:00:30 AM10/29/19
to beast...@googlegroups.com
The checkpointing scheme doesn't append to existing log files to minimise the risk of data loss. The suggestion is to use the '-prefix' option to add a prefix to all the log filenames to make it clear it is the restart of a checkpoint. Then collect and append the logfiles using log combiner after the run is finished.

Andrew

--
You received this message because you are subscribed to the Google Groups "beast-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beast-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/beast-users/379cd759-6211-4258-a03d-633a64632164%40googlegroups.com.

Luiz JQ

unread,
Oct 29, 2019, 6:39:14 AM10/29/19
to beast-users
Thanks, Andrew!
Reply all
Reply to author
Forward
0 new messages