[tahoe-dev] best practice for wanting to setup multiple tahoe instances on a single node

[Jimmy Tang]

On Wed, Jan 11, 2012 at 2:03 AM, Vladimir Arseniev <vladimira at aport.ru> wrote:
> On 12-01-11 04:11 AM, Jimmy Tang wrote:
>
>> ... I have 150tb of space on a machine but I can only make
>> a bunch of 16tb partitions.
>
> What's the OS? Could you aggregate those bits using software RAID, LVM,
> xfs or whatever?

Right now it's rhel6, the problem is filesystem limits and our
laziness and dislike of solaris as it doesn't conform with the rest of
our systems, having something different from what we already run means
more work for us :) xfs has a limit of 32tb in practice, I think the
last time I tried making big fs's most of them had a 16tb limit due to
the blocksizes they used (ext3,4,xfs), there are also problems with
fsck's large systems which has also put me off a bit. I haven't tried
jfs partly cause I am lazy with building the modules. We're still
evaluating this, btw we have 2x135tb and a few other 50tb sized
systems that we're thinking of spreading data out, if it works we
might just get a few more of these 135tb units and build up the
system.

Jimmy.



-- 
http://www.sgenomics.org/~jtang/