Talk:GPFS: Difference between revisions

Content deleted Content added

Inline

Revision as of 16:17, 30 March 2011

Computing: Software Start‑class Mid‑importance

	This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.ComputingWikipedia:WikiProject ComputingTemplate:WikiProject ComputingComputing
Start	This article has been rated as Start-class on Wikipedia's content assessment scale.
Mid	This article has been rated as Mid-importance on the project's importance scale.
	This article is supported by WikiProject Software.

GPFS 3.3

IBM demoed a pre-release of GPFS 3.3 during SuperComputing 2008 (November 2008) with full support of Windows HPC Server 2008

http://www.youtube.com/watch?v=mqimygrrHTw&feature=related —Preceding unsigned comment added by 81.165.2.78 (talk) 08:53, 3 February 2009 (UTC)[reply]

Removed the 'Advert' tag. If someone disagrees, please indicate here which statements are problematic. --Dan.tsafrir (talk) 14:44, 19 May 2009 (UTC)[reply]

Need to Compare Against ZFS?

The article compares GPFS to Google GFS and Hadoop HDFS, but it seems that ZFS might be more similar to GPFS than the latter two, and so a GPFS vs. ZFS comparison could be more appropriate / helpful. --Dan.tsafrir (talk) 14:53, 19 May 2009 (UTC)[reply]

Not so sure. The only GPFS installation I've seen was on a SAN attached to a supercomputer cluster, where the ability to stripe data across every disk meant that when you asked for a file, you got every disk head fetching a bit of it, then the san bandwidth bringing it to you. Does its a premium alternative to things like HDFS, which has worse remote bandwidth but does work near the data instead. I don't know how ZFS stands up to either use. It may scale, but does it have the bandwidth or the locality? (COI disclaimer, I work on hadoop clustering) SteveLoughran (talk) 21:47, 11 June 2009 (UTC)[reply]

File name length

The info box states »256 UTF-8« as the maximum file name length which is pretty nonsensical. Does this mean »256 bytes in which an UTF-8-encoded name is stored« (making this essentially »256 UTF-8 code units«) or »256 Unicode code points«? The table at Comparison of file systems is even worse, stating »255 UTF-8 codepoints« which (a) deviates from the measure given here and (b) is even more nonsense, as there is no such thing as an »UTF-8 code point«. I wasn't able to find clarification on IBM's web site which seems to be pretty silent on the limits and restrictions of the file system in the general case. Maybe someone is able to clarify this, as how it currently stands it's a useless piece of information (or rather,not information at all). —Johannes Rössel (talk) 12:34, 7 July 2010 (UTC)[reply]

Shared Disk

The article states (in the very first sentence) that GPFS is a shared-disk network filesystem. While thats certainly the common case for GPFS, I have two gpfs filesystems that have no shared disk. Since at least 3.0, gpfs supports storage local to only one storage node. It even has features for 'failure domains' to be able to replicate around a problem affecting a single node.

However, I'm new to wikipedia and dont' want to edit the main article right away. Whats the best way to update this page?