[olug] NAS Cluster

[Carl Lundstedt]

hadoop?
http://hadoop.apache.org/core/
We're testing hdfs at our shop as a replacement for dcache (dcache.org) 
which just stinks.
We have 200+ TB in hadoop and all our 'small' pools (<1TB) have worked 
great, all our medium pools (3TB<pool<10TB) have had some issues, but 
mostly hardware.  I'll be putting in one of our thumpers (48TB) next 
week to see how it works in hadoop.

Hadoop seems to be much more fault tolerant than other distributed 
filesystems we've used.


Carl Lundstedt
UNL
> Hello,
>       Anyone doing NAS clustering around here?
> I will have 3 pallets of new servers in a few weeks, and I get to use some
> of them, perhaps 12 or more for experimentation.
>
> My goals are open source, cost, performance and redundancy but mainly
> redundancy.
>
> I've seen it done with iSCSI and LVM, but i havnt seen anyone pull the plug
> on a node and be able to keep the data intact.
>
> I've been reading up on PVFS2, and that looks promising.
>
> Just seeing if anyone has done this with great success, I'd like to be able
> to pull the plug on a node and not lose any data.
>
> Machine specs are Dell R200, quadcore intel, 2GB ram, 146GB raid1, dual 1GbE
> I need to find some more switches now
>
> Thanks and have a great weekend ppl
> _______________________________________________
> OLUG mailing list
> OLUG at olug.org
> https://lists.olug.org/mailman/listinfo/olug
>