Backup help
David Cousens
davidcousens at bigpond.com
Wed Sep 20 17:19:23 EDT 2017
On Wed, 2017-09-20 at 06:31 -0700, John Ralls wrote:
>
> >
> > On Sep 20, 2017, at 1:35 AM, DaveC49 <davidcousens at bigpond.com>
> > wrote:
> >
> > Hi Colin,
> >
> > i agree with both you and Bram there. It is not a reliable backup
> > of the
> > database but really only useful for sharing a database across
> > several
> > machines and worked well for the use case where my laptop was
> > generally
> > connected to my LAN and I worked away from home
> > occasionally. haven't had
> > any experience with how it copes with simultaneous access and
> > locking
> > issues. There is also the MYSQL Cluster as another approach. You
> > also need
> > to check that the replication is up to date and completed before
> > assuming
> > the databases are identical. I use a cron job and mysqldump to
> > dump the
> > databases to an NAS for backup.
> Um, wouldn’t it be easier to just use one of the machines as a server
> and connect to it with GnuCash from the other machines?
John,
It is easier for the general use case where the two computers remain
internet connected at all times to just use a server.
Like Art I set it up half to explore database replication and half to
solve a problem of being able to just pick up my laptop and go. I
prefer to develop in a desktop environment but had the need to demo app
development to clients away from home without easy internet access and
sometimes continue development while I was away from home. I did not
have my desktop setup as a server and for various reasons didn't want
to. Using replication to sync the databases meant I could just pickup
the laptop and go without having to copy database files from one
machine to the other specifically. I also keep my home directories
synced across the two machine.
With the development of mobile apps with in built databases with a need
to be synced with a central database but where internet access is not
continuous but database access is required when not connected there is
a perfect use case. There are collision problems with multiple users
particularly and as replication is asynchronous the database copy can
get out of sync. I can't claim any great expertise but there are people
addressing those problems.
The main beauty of it is that it runs totally independently of any app
accessing the database. The distributed database aspect is also
interesting as any failed copy can be restored from any still working
copies giving the RAID like redundancy mentioned earlier in the thread
>
> Regards,
> John Ralls
>
More information about the gnucash-user
mailing list