[DRBD-user] Is `sync` option of ext3 mount SHOULD used in drbd HA clusters?

[Pan'ko Alexandr]

В Пнд, 15/09/2008 в 15:26 +0200, Lars Ellenberg пишет:
> On Mon, Sep 15, 2008 at 03:55:55PM +0300, Pan'ko Alexandr wrote:
> > Hi.
> > 
> > Let's see article from 
> > http://wiki.centos.org/HowTos/Ha-Drbd
> > 
> > Let's do commands:
> > 
> > [root at node1 /]# drbdadm primary repdata ; mount /dev/drbd0 /repdata
> > [root at node1 /]# dd if=/dev/urandom of=/repdata/test.file  bs=1000000c
> > count=1000; ip l set eth0 down;
> 
> 
> how about
>   dd ... conv=fsync
> 
> see also: man fsync

Of cause I now about this params.
But are you sure that Tomcat and Firebird know about it too? :)

By such commands I modeling the accident fail of one of peer in active
disk writes.

I know how to save files ;).

> > [root at node1 /]# ls -l /repdata/test.file
> > -rw-r--r-- 1 user users 1000000000 Sep 15 15:07 /repdata/test.file
> 
> thats all in page cache still.
> 
> > [root at node2 /]# drbdadm primary repdata ; mount /dev/drbd0 /repdata
> > [root at node2 /]# ls -l /repdata/test.file
> > -rw-r--r-- 1 user users 868456004 Sep 15 15:07 /repdata/test.file
> > 
> > As you will see there will be different sizes.
> > But they SHOULD be the same.
> > 
> > If we will add to mount command the option '-o sync', than there will
> > be no differences in sizes. But in case of 'bs=1c' in dd command the
> > file creation will be very and very slow because of turned off chaching.
> > 
> > Anybody knows this issue?
> > SHOULD I turn off caching om ALL underling devices and filesystems to
> > get HA cluster filesystem?
> 
-- 
Regards, Alexandr A. Panko