Robert, On Thu, May 23, 2013 at 10:04:29AM -0600, Charles Curley wrote: > On Thu, 23 May 2013 13:59:33 +0000 > "McGraw, Robert P" <rmcg...@purdue.edu> wrote: > > > Why does amanda stop at %52 when I still have 1.5TB of data in the > > holding disk to write to the tape? It is hard to believe that the > > LTO4 compression is so bad that I am not getting any compression at > > all. > > It would be if you have both tape drive compression turned on (as you > do) and compression specified in the disk list entries (DLEs). > Conventional wisdom on this list has been that DLE compression (using > gzip, say), is more efficient than tape drive compression.
Tape drive compression is nice, and the newer technologies do a block by block determination as to whether the data can be compressed or not [used to be that if you SW compressed and then HW compressed you might get inflation and not only waste CPU but end up occupying more tape than if you hand't compressed at all]. What is more difficult is estimation of tape usage. SW compression gives you a raw size number, a compressed size number and Amanda can sum the compressed size numbers up and compare them to the capacity of the tape. With HW compression its kid of out of your hands and you don't really know, except from experience, what is going to fit on tape any given night. If you run SW compression and see you get 1/3 data reduction, you can move the DLE to HW compression and will hopefully get 1/3 data compression, which you might allow for by "lenthening" the tape capacity by 1/3 the size of the particular DLE. At least, I've done this in the past. I don't currently run HW compression on any DLE or amanda config that I'm managing. YMMV. Database compression depends a lot on your database. I have a lot of sparse databases that compress very well. This is not always the case, you have to know your data. > -- > > Charles Curley /"\ ASCII Ribbon Campaign > Looking for fine software \ / Respect for open standards > and/or writing? X No HTML/RTF in email > http://www.charlescurley.com / \ No M$ Word docs in email > > Key fingerprint = CE5C 6645 A45A 64E4 94C0 809C FFF6 4C48 4ECD DFDB --- Brian R Cuttler brian.cutt...@wadsworth.org Computer Systems Support (v) 518 486-1697 Wadsworth Center (f) 518 473-6384 NYS Department of Health Help Desk 518 473-0773
