best file system for data server

Asked by 4heid

i just rescued my reiserfs drive with the incredible help of this board.
however, now that we will be redoing our data server using an ubuntu box instead of a standalone drive unit, what is the ideal file structure to use for this secondary drive that will run off of Ubuntu as a slave drive and will most lilkely have a Raid1 backup as well.

Question information

Language:
English Edit question
Status:
Answered
For:
Ubuntu util-linux Edit question
Assignee:
No assignee Edit question
Last query:
Last reply:
Revision history for this message
Bhavani Shankar (bhavi) said :
#1

Unfortunately, as far as I know, the answers to the performance and scalability questions may be different. ReiserFS claims to handle small files exceptionally well, and (if they're very small) it can be very efficient in terms of disk space consumption. On the other hand, XFS is supposed to be an industrial-strength, highly scalable filesystem, with astronomical upper limits in terms of file counts, sizes, and so on. The other thing to note about XFS is that it's (comparatively speaking) very RAM-hungry. Depending on your point of view, that could be very good, or very bad..

There are lots and lots of benchmarks out there that compare these (and often some others, like JFS, for instance) against Ext2/3.

When it comes to actually using them, the two major things to note about ReiserFS and XFS is (a) that their approach to filesystem integrity protection is quite different, and (b) that running them over RAID (which I presume you will be doing) requires careful consideration, to avoid taking a performance hit.

 it's worth taking a look at a comparison of their design limitations ... workable block sizes, max file counts, RAID-friendliness, max partition sizes, etc. (Wikipedia should be able to help you with that kind of cursory comparison.)

Regards

Can you help with this problem?

Provide an answer of your own, or ask 4heid for more information if necessary.

To post a message you must log in.