ZFS zdb -S ... what means the first column?

RM.

RM.

Hello,

for example:
Code: 
#zdb -S tank

Simulated DDT histogram:

bucket                 allocated                      referenced   
refcnt  blocks  LSIZE  PSIZE  DSIZE  blocks  LSIZE  PSIZE  DSIZE
------  ------  -----  -----  -----  ------  -----  -----  -----
  1  69.5M  8.66T  8.45T  8.46T  69.5M  8.66T  8.45T  8.46T
  2  1.63M  207G  194G  195G  3.32M  423G  396G  398G
  4  21.7K  1.95G  1.44G  1.53G  100K  9.33G  6.96G  7.34G
  8  2.70K  305M  213M  223M  26.1K  2.87G  2.05G  2.14G
  16  1.38K  164M  140M  142M  28.3K  3.28G  2.87G  2.91G
  32  91  4.69M  2.79M  3.30M  3.65K  183M  102M  124M
  64  54  3.44M  1.31M  1.66M  4.88K  318M  123M  155M
  128  99  10.2M  2.64M  3.21M  19.9K  2.10G  546M  662M
  256  3  384K  144K  152K  1.18K  152M  64.9M  68.0M
  512  3  256K  2.50K  24.0K  2.34K  198M  1.94M  18.7M
  1K  2  128K  1.50K  16.0K  2.75K  212M  2.20M  22.0M
question:
What mean's the column "refcnt"? Especially the row "Refcnt=512" or "=1K"? Any idea's? :)
 
For bucket refcnt=512, the block count is 3. This means there are (or would be) three blocks which are referenced at least 512 times AND less than 1024 times.
 
Hello,

in other words there are '69.5M blocks' physically allocated, which are only onetime logically referenced and have no other duplicates on disk. So all rows with 'refcnt > 1' are blocks, which can be deduplicated. Is this right?
 
You must log in or register to reply here.
Back
Top