Date: Wed, 07 Aug 2002 12:57:16 -0500 From: "Jack L. Stone" <jackstone@sage-one.net> To: "Siegbert Baude" <Siegbert.Baude@gmx.de>, "Rob Ellis" <rob@web.ca> Cc: <freebsd-questions@FreeBSD.ORG>, =?Windows-1252?Q?S=F8ren_Schmidt?= <sos@freebsd.dk> Subject: Re: Best "bs" for dd copies (was: Re: Questions about vinum and failure of root partition) Message-ID: <3.0.5.32.20020807125716.02d62db8@mail.sage-one.net> In-Reply-To: <001c01c23e25$b05a1180$406a3c86@whwurm.uniulm.de> References: <005e01c23dcb$061acbb0$6602a8c0@swbell.net> <200208070101.g7711iU06306@clunix.cl.msu.edu> <005e01c23dcb$061acbb0$6602a8c0@swbell.net> <3.0.5.32.20020807085441.02d62db8@mail.sage-one.net>
next in thread | previous in thread | raw e-mail | index | archive | help
At 05:18 PM 8.7.2002 +0200, Siegbert Baude wrote: >Hi Jack, > >first an apology to Soeren for cc'ing him, but he probably is the best >man to answer this question (and a second one for not being able to >produce the correct letter for the spelling of his name *g*). > >Rob wrote: >> >then create a backup of the fbsd partitions on >> >the first disk, copying everything from ad0s1 to ad2s1: >> > >> > dd if=/dev/ad0s1 of=/dev/ad2s1 bs=102400 > >Jack wrote: >> My side question is about the "dd" command. Why did you choose the >> parameter "bs=102400" rather than any other?? I've been using 8192, >but >> have seen this switch all over the map, including 1024 to 1M. I know >it can >> make a difference in the time to do an image because of the sizing. >With >> 8192, I do an entire 40GB HD in 39 mins (1.4GHz CPU) but takes 49 mins >for >> a 1GHz CPU. > >I just experimented a bit last weekend, when I dd'ed a 80GB IBM >IC35L080AVVA07-0 to another disk of exactly the same type (needed a >bit-copy as backup, because the partition table was corrupted). > >I tried bs from standard 512 up to 8MB by always doubling the value from >try to try and found that 128k worked the best for me. That is quite >near to the proposed 100k of Rob. The transfer rate varied from 6 MB/s >to 20MB/s, if my memory works right. At least it was far away from the >33MB/s the UDMA-33 mode should give (the disks could do UDMA 100 and the >highpoint controller even UDMA 133, but I only had normal cables). IBM >claims its disk should transfer a sustained rate from 48MB/s to 23MB/s >depending on the zone. The disk cache is 2MB, btw. The board was an >Abit-BX133 with 256MB RAM and a PIII-850. > >So, if anybody knows how to calculate the best value out of the >technical parameters or can explain, why ~100k seems the best value (and >not e.g. something in the area of disk cache size) I also would be very >interested to hear. >What is the maximum at all, one can expect? Is it possible to reach the >maximum rate IBM claims for its disk with dd? > >Ciao >Siegbert > As a follow-up, I adjusted the bs=8192 to bs=102400 and the following happened: 1) time to backup reduced by 50% 2) CPU usage reduced by 90% Pretty dramatic effect. Wonder what it does to the integrity? But, less time should mean less file changes within the duration of the backup....?? Best regards, Jack L. Stone, Administrator SageOne Net http://www.sage-one.net jackstone@sage-one.net To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3.0.5.32.20020807125716.02d62db8>