[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
'rm -rf' alternatives
- To: misc_(_at_)_openbsd_(_dot_)_org
- Subject: 'rm -rf' alternatives
- From: Darik Horn <darik_(_at_)_gulf_(_dot_)_waterways_(_dot_)_geeky_(_dot_)_net>
- Date: Tue, 19 Sep 2000 18:24:27 -0400 (EDT)
(Hi. I'm a new OpenBSD user that has been lurking for a short while.)
Some UFS operations are painfully slow. For example, running 'rm -rf'
against '/usr/ports' on a Pentium/150 takes several minutes. Doing the
same thing on my 486/66 toybox, even with a SCSI disk, is a coffee job.
- Is it poor form to use 'rm -rf' on a BSD box? Is there some better
alternative? Is UFS just slow in this peculiar example? (I'm thinking
that the OpenBSD system is just extraordinarily conservative -- a good
thing -- as compared to the other systems that I would normally use.)
- Enabling softupdates did not noticeably improve performance. What about
an async mount? I would never use async on an important partition, but is
async suitable for deskbox use or sandbox use? Rather, do normal users
like yourself use async, or is it just an option for the masochistic?
- What are the usual modes of failure for a botched async update? Does
the whole filesystem get hosed, or will I just get some e2fs-esque fsck
nagging during reboot?
- Is the filesystem sluggish only on slower/older/smaller hardware? Say,
machines with less than 128 megs of system memory? (I'm not yet
comfortable enough with OpenBSD to put it on beefy hardware, and
OpenBSD-2.7 does not run well under VMware.)