Date: Fri, 3 May 1996 14:34:01 -0700 (MST) From: Don Yuniskis <dgy@rtd.com> To: wilko@yedi.iaf.nl (Wilko Bulte) Cc: freebsd-hackers@freefall.FreeBSD.org (FreeBSD hackers) Subject: Re: using DLT drive on FreeBSD Message-ID: <199605032134.OAA09163@seagull.rtd.com> In-Reply-To: <199605031845.UAA01588@yedi.iaf.nl> from "Wilko Bulte" at May 3, 96 08:45:24 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> > Don't count on it! My 2500xt (15/30G) is already starting to max out > > when doing level 0 dumps.... :> Plus, the tapes are quite expensive > > and DON'T EVER DROP ONE! > > Hmm, I don't have 10Gb worth of disks at home ;-) Tapes are a bit Give it time... I'm *sure* you WILL! ;-) > of an issue, in supporting customers I learned you can screw the > tapes. Never succeeded in botching one myself but as we all know > customers tend to be better at breaking things. Yes, they really don't like to be dropped. If the spindle gets bound up inside it (i.e. so the tape can't move), you can break the "tape extractor" in the drive (major bummer). > > I wouldn't even try to keep them streaming -- unless you've got a really > > fast disk subsystem *and* the DLT on a separate (FAST) SCSI adapter > > with nice short cables... > > Come on, they are 'Born to Stream' ;-) The Sun could do it, so I don't Heh heh heh... yeah, just getting the host to cooperate is the big problem! :> My comment was intended kinda "tongue-in-cheek"... more to say "Argh! These things are so damn fast that the processor becomes the bottleneck... be thankful and live with it!" ;-) > see why a decent FreeBSD system could not do it. FWIW the single disk > I pulled data from on the Sun is an identical model to the one in my > Asus @ home. I assume you've got everything set up for synchronous transfers, etc. (otherwise, most async SCSI stuff maxes out around 4MB/sec -- with typical cabling -- which would cut it close if disk and tape are on same bus) > A controllerbased raidarray on the Sun proved to be better in getting > the drive to stream but it is not impossible to do it from a single disk. I imagine some of the SCSI system implementors might be better at explaining the possible bottlenecks. BTW, there are some resources available (I think on Quantum's page or maybe ftp site, etc.) that you might want to grab. They allow firmware updates to the drive, etc. Don't know if they work with the DEC 2000 product or just Quantums stuff, etc. (Quantum bought the product line from DEC and I don't know how the support is handled) --don
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199605032134.OAA09163>