wyatt@cfa.UUCP (Bill Wyatt) (03/13/86)
I am trying to find out something about what people have done on disk striping,
that is, increasing I/O by distributing a file across more than one disk
spindle. I am interested about such work in both the Unix and VMS domains.
If there is sufficient interest, I will, of course, summarize to the net.
Thanks.
Bill UUCP: {seismo|ihnp4|cmcl2}!harvard!talcott!cfa!wyatt
Wyatt ARPA: wyatt%cfa.UUCP@harvard.HARVARD.EDU
--
Bill UUCP: {seismo|ihnp4|cmcl2}!harvard!talcott!cfa!wyatt
Wyatt ARPA: wyatt%cfa.UUCP@harvard.HARVARD.EDUhosking@convexs.UUCP (03/17/86)
All currently supported versions of Convex's port of 4.2 BSD support disk
striping. We also support basic disk block sizes of up to 64 KB, and some
asynch file I/O features. These can be used together for an even greater
performance boost. 99.9% of all user programs don't know or care that
striping is being used. Except for the increased performance, striping is
virtually transparent to the user, too. (There are a few trivial changes to
fstab, and a new file called /etc/stripecap, but this is no big deal.)
Striped file systems are built using a utility called newst, which is very
similar to newfs, but builds striped file systems. Newst allows you to
stripe together any reasonable combination of partitions. (The root
partition can't be striped, for technical reasons.) Once a striped
partition is built with newst, it is mounted as /dev/stxx instead of
/dev/daxx, using a different major device number. It can then be
accessed just like any "normal" partition.
Another, less obvious feature of striping is that it gives you increased
flexibility with partition sizes. If you have several "a" partitions that
are too small to be useful, you can stripe them together to form a larger
partition, without the need to edit kernel tables, recompile, etc.
I really don't want to turn this into a commercial for Convex. If you have
questions, I'll be happy to try to answer them and/or refer you to our
sales people.
Doug Hosking
Convex Computer Corp.
Richardson TX
{allegra, ihnp4, uiucdcs}!convex!hoskingjack@boring.uucp (Jack Jansen) (03/21/86)
Pardon my ignorance, but what is so special about striping? It seems to me that interleaved filesystems will automatically give you striping, provided you choose your m and n intelligently, and they've been in unix since V6, at least. Do I miss something? -- Jack Jansen, jack@mcvax.UUCP The shell is my oyster.
andrew@aimmi.UUCP (03/26/86)
In article <27300003@convexs> hosking@convexs.UUCP writes: > >..... 99.9% of all user programs don't know or care that >striping is being used. Except for the increased performance, striping is >virtually transparent to the user, too. >accessed just like any "normal" partition. > What is disk striping?? Interleave, I can cope with... but striping?? >Another, less obvious feature of striping is that it gives you increased >flexibility with partition sizes. If you have several "a" partitions that >are too small to be useful, you can stripe them together to form a larger >partition, without the need to edit kernel tables, recompile, etc. Now, that sounds *handy*. -- ------------------------------------------- Andrew Stewart USENET: ...!mcvax!ukc!aimmi!andrew "My axioms just fell into a Klein bottle"
charisse@convexs.UUCP (03/26/86)
Striping refers to the capability of spreading a file system across multiple disks. It allows you to take advantage of parallel I/O operations at the hardware level and can increase I/O throughput. For example if you have 4 disks you can intiate 4 reads at once. Thus increasing throughput by approximately 4X. Of course you still have to pay the piper. Your I/O subsystem has to have knowledge of the disk layout in order to optimally spread your file across multiple disks. And what do you do if you are striping together disks with partitions of unequal size? Both the Cray and the C1 provide a disk striping capability. It seems to be a win for large machines. charisse ihnp4!convex!charisse convex!charisse@rice.arpa