[DRBD-user] primary/primary with ext2 not updating real-time
theophanis_kontogiannis at yahoo.gr
Sun Feb 24 23:26:37 CET 2008
Thank you for your time and answer,
You are right no one would want ext2, not even me. It was a typo.
I meant ext3 :)
Now regarding what I try to do.
When I start to write the file "ABC" on server 1, and keep on doing "ls" on
server 2, the file does not appear to be there. Not even after the file is
written completely on server 1, it still does not appear on server 2. It
will only appear on server 2, if I un-mount and re-mount the volume.
The same applies vice versa on a file that is written on server 2 and
looking for it on server 1.
It looks like the update on the file system, is not real time.
However I do not know if this is an issue of the block device (i.e. drbd
running in primary/primary mode, so there is no secondary to have it synced)
or the file system itself (and here fits in the cluster aware file system,
From: devnull at deleted.on.request.example [mailto:devnull at deleted.on.request.example] On Behalf Of someone
Sent: Sunday, February 24, 2008 10:13 PM
To: Theophanis Kontogiannis
Cc: drbd-user at lists.linbit.com
Subject: Re: [DRBD-user] primary/primary with ext2 not updating real-time
Theophanis Kontogiannis wrote:
> Then when I write file "ABC" on server 1, I need server 2 to realize
> (with "ls") that file "ABC" is being written (or at least was written)
> WITHOUT having to un-mount and mount again.
Uhm, while I fear I don't really understood what you are doing in your
setup.. you know you need to run a cluster aware file system on any sort
of shared block device -- be it DRBD, ENBD, iSCSI or something else and
that ext2 is clearly not one of those? Look into OCFS or GFS. Besides
that it's beyond me why anyone would like to use ext2 these days. If
what you are after is a parallel file system, I'd recommend to look into
GLusterFS with AFR translator.
More information about the drbd-user