elb at pidgin.im
Fri Nov 16 21:32:52 EST 2007
Kevin Stange spake unto us the following wisdom:
> Ethan Blanton wrote:
> > I am not at *all* sure that using the journal as a lockfile is safe.
> > (It may be, I am not saying it certainly is not.) 270 seconds is a
> > reasonably conservative pull timeout.
> From my investigation, monotone does the whole pull as a SQL transaction
> and finishes with a commit, so the transaction seems to be entirely
> contained within the journal until the commit is issued -- no mtime
> change for the main DB file until after that.
My understanding from the monotone mailing list and IRC is that
netsyncs are checkpointed from time to time, so that if a pull is
aborted partway through, it does not (always) have to be restarted
from the beginning. I believe there are triggers on both number of
revisions and total amount of data transferred, though I am not sure.
The important data point, though, is that whether or not this
correlates with the presence of a journal file is dependent on
sqlite's internal operation, and not something we should depend on.
Is there a reason that /usr/bin/lockfile is not appropriate?
The laws that forbid the carrying of arms are laws [that have no remedy
for evils]. They disarm only those who are neither inclined nor
determined to commit crimes.
-- Cesare Beccaria, "On Crimes and Punishments", 1764
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 189 bytes
Desc: Digital signature
More information about the Devel