Date: Mon, 15 Jul 2002 19:21:29 -0700 From: Nick Sayer <nsayer@quack.kfu.com> To: "Tortise@Paradise" <tortise@paradise.net.nz> Cc: freebsd-stable@freebsd.org Subject: Re: Poor Mans Software raid 1 on root partition? Message-ID: <3D338329.4010205@quack.kfu.com> References: <25f401c228d4$a3482fb0$1a01000a@area51><20020711091015.B51520@flake.decibel.org><20020711200902.3653b534.steve@sohara.org><20020713032546.GD61459@wantadilla.lemis.com><20020713075109.06ecf02f.steve@sohara.org><20020713100218.B284@twincat.vladsempire.net><20020713222745.00281f72.steve@sohara.org><20020714004247.GB16279@wantadilla.lemis.com> <20020714071524.1587f419.steve@sohara.org> <007801c22c6c$9dc32fe0$0900a8c0@P1200n>
next in thread | previous in thread | raw e-mail | index | archive | help
It would be far, far better to simply do a daily backup using 'dump'. For just a little more money than buying a whole 'nother drive, you could buy a tape drive. Tape drives are really ideal for backups because you can store the media other than inside the machine you're backing up. :-) If you don't want a tape drive, it's still a better idea to do dumps to another disk rather than dd. Why? 1. You can compress the output of dump. 2. You don't have to have identically partitioned disks 3. You won't get a "clean" filesystem copying it with dd. In fact, the source filesystem may change enough over the course of the dd that the resulting image may be completely useless. 4. You can potentially store many, many days worth of backups on the alternate drive. This lets you restore not just last night's backup, but potentially last *week's*. 5. You can use the 'nodump' flag (see chflags(1)) to exclude files from the backup that are uninteresting. Yes, doing a restore takes a little more time than simply booting the other drive. But in practice, the likelyhood that you will really *need* to do so is sufficiently low as to not be worthwhile, IMHO. The only reason to consider software raid for a truly mission critical application is if you've got a really large dataset in an external box that you can move from one machine to another if you need to get back up quickly. In that circumstance, presumably the contents of the system disk of the machine don't matter, meaning that the application would come back on line simply by moving the disk to another machine and restarting it. If that's unreasonable, then the whole machine requires RAID (among other things), which means you'll be getting a hardware RAID. But don't ever forget that RAID won't help you if you accidently do an rm -rf / as root. :-) Data integrity is NOT a substitute for good backups. Tortise@Paradise wrote: > Reading this thread it seems that this is currently not an option. > > However why can't we do a poor mans RAID, namely run a cron job say daily > (or whenever) which bulk copies / updates one (complete) HDD onto a second > one, be it SCSI or IDE, with a view that should the main HDD fail the second > can still be booted from and the only data loss will be the interim since > the last "backup". Comments on this strategy would be appreciated. (No! I > do not work for a HDD manufacturer.....LOL) > > David Hingston MB ChB MBA > _________________________________________________________________________ > tortoise@paradise.net.nz > http://hingston.yi.org/ > http://pcmc.yi.org/ > If you seek a digitally signed response please advise. > If you received a warning on reading this e-mail, please go to > http://www.baycorpid.com/settings/email.asp?CA=healthcert to update your > settings > > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-stable" in the body of the message To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3D338329.4010205>