[rescue] Block size and the single DD - more test results

Sheldon T. Hall shel at cmhcsys.com
Thu Feb 12 15:44:01 CST 2004

 Sheldon T. Hall wrote, inter alia ...
> It looks like the disk can reasd the /var filesystem in as little as 29
> seconds, but the tape transfer process takes at least 1:12 to write that
> data.  That's 64701KB/29sec = 2231 KBs reading, 64701Kb/72sec = 898 KBs
> writing.  I assume a constant part of the write time is set-up and
> tear-down, so the per-second speed on a larger filesystem would be higher.

I predicated that on dd being smart enough to read only the used part of the
slice, i.e. the actual data ... but in running the same tests on the SGI
Challenge L, it looks like dd reads the _whole_ slice.

If that's so, then the figures are ...

	read	122887KB/29sec = 4,237 KBs
	write 122887KB/72sec = 1,706 KBs

which seems more reasonable, non?  IN both cases I was reading from
/dev/rdsk/xxxxx, i.e. the raw device.

However, if dd reads the whole slice, it would seem to be a poor choice of
backups of small filesystems on big disks/slices ... yet many folks use it.

Anyone want to tell me what's going on here?


More information about the rescue mailing list