[mylvmbackup] Re : Re: mylvmbackup
- From: Pascal Borghino <pborghino@xxxxxxxx>
- To: mylvmbackup@xxxxxxxxxxxxx
- Date: Wed, 2 Apr 2008 09:03:25 +0000 (GMT)
Hi,
I do not know if you hit the bug already,
https://bugzilla.redhat.com/show_bug.cgi?id=204791 ( quite old but still )
> LVM snapshot on RHEL AS 4 U4, Linux 2.6.9-42.ELsmp x86_64, 64-bit.
> create a snapshot with size 32M
> Mount the snapshot volume.
> copy a 64M file to the original volume...
> The copies complete and the snapshot become full. Umount the snapshot vol and
> try to remove the snapshot using lvmremove.
> At this point, it crashed the server...
For flushing,
>First that mylvmbackup includes an extra_flush_tables to workaround
>interactive timeouts. Is this referring to the mysql client
>interactive timeout? Wouldn't it be better to have a mylvmbackup option to
>adjust the interactive timeout instead of making the flush
>(and thus the interference in mysql) take longer?
sounds good.... as long we are above MySQL 4.0 :) yes... do not ask ;)
>Have you benchmarked the flush each table then flush tables with read lock
>method to see if it's faster? Internally any updates that >were waiting while
>the flush tables per table is being ran will try to run and open those tables
>up again potentially make the flush tables >with read lock wait until they are
>done running so it an grab it's locks and flush all the tables.
I did not benchmark it yet.. it is in the todo... might be handy knowing the
behavior of my application.
About version control and hosting, I also heard about mercurial... but
basically I have no preference...
Cheers
P.
_____________________________________________________________________________
Envoyez avec Yahoo! Mail. Plus de moyens pour rester en contact.
http://mail.yahoo.fr
Other related posts:
- » [mylvmbackup] Re : Re: mylvmbackup