Date: Wed, 02 Oct 1996 10:03:12 -0700 From: James Graham <greywolf@siva.captech.com> To: "Kevin P. Neal" <kpneal@pobox.com> Cc: hackers@freebsd.org, tech@openbsd.org, current-users@netbsd.org Subject: Re: VPS mailing list, BSD interest? Message-ID: <199610021703.KAA04834@siva> In-Reply-To: Your message of "Mon, 30 Sep 1996 18:46:08 EDT." <1.5.4.32.19960930224608.0066dee0@mindspring.com>
next in thread | previous in thread | raw e-mail | index | archive | help
"Kevin P. Neal" sez: [much deleted] # # It's a rather long document, and it gets into much detail in some places. # # Neat stuff. Resizable partitions, easy mirroring of individual partitions, # reorganizing of partitions with ease. Stability in the event of hardware # going bad in run time. Imagine taking a system down, hooking up a drive, # then booting multi-user mode. In multi-user mode, add partitions and # filesystems as you like. Rearrange as you like. Minimized down time. # # "Help! My home directory partition is full!" # # "Hold on a sec.....How's that?" # # "Wow! Thanks!" Yes, this is cool, I agree. Now, one thing that none of the other filesystems deal with is how to *shrink* a partition, should it become necessary. I agree that it would be a cool thing to grow/shrink/move filesystems. Someone wrote that an extensible filesystem would be difficult because of the overhead of layout, as well as the problem of fragmentation. I note that fsck has a '-d' option for 'debug'; perhaps we should implement a -D option for "defragment". I know defragmenters can take a LOT of time, but if extensible filesystems can't avoid fragmentation easily, perhaps a FS_FRAGMENT flag should be set in the filesystem superblock, and fsck would pick it up and attempt to defrag the filesystem at the next reboot (really fsck -p). But I still have my reservations about keeping lots of arcane metadata in obscure locations. If this could all be made to work with a maximum of transparency, I'll probably shut up and not bitch too much (not that it matters as my technical contributions have been minimal:-). [It's frustrating being able to half-grasp the concepts being bandied about and not being able to a) find time to write code and b) understand the code to be written.] (stream_t *) (--*greywolf)->u_thought = 0;
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199610021703.KAA04834>