I'm doing a putback onto my local workstation, watching the disk
activity with "zpool iostat", when I start to notice something
quite strange...
zpool iostat 1
capacity operations bandwidth
pool used avail read write read write
---------- ----- ----- ----- ----- ----- -----
biscuit 3.39G 185G 0 4 4.64K 39.1K
biscuit 3.39G 185G 0 0 0 0
biscuit 3.39G 185G 0 118 0 1.16M
biscuit 3.39G 185G 0 0 0 0
biscuit 3.39G 185G 0 0 0 0
biscuit 3.39G 185G 0 0 0 0
biscuit 3.39G 185G 0 0 0 0
...
biscuit 3.39G 185G 0 113 2.47K 1.28M
biscuit 3.39G 185G 0 0 0 3.96K
biscuit 3.39G 185G 0 0 0 3.96K
biscuit 3.39G 185G 0 0 0 3.96K
biscuit 3.39G 185G 0 4 0 360K
biscuit 3.39G 185G 0 109 0 699K
biscuit 3.39G 185G 0 0 0 0
biscuit 3.39G 185G 0 0 0 0
biscuit 3.39G 185G 0 0 0 0
biscuit 3.39G 185G 0 0 0 0
biscuit 3.39G 185G 0 137 0 3.26M
biscuit 3.39G 185G 0 2 0 380K
biscuit 3.39G 185G 0 0 0 127K
biscuit 3.39G 185G 0 0 0 127K
biscuit 3.39G 185G 0 2 0 71.3K
biscuit 3.39G 185G 0 113 1013 682K
biscuit 3.39G 185G 0 0 0 3.96K
biscuit 3.39G 185G 0 0 0 0
It claims there are no read operations but the read
bandwidth is non-zero. Is this true or a bug? If
it is true, then how is it possible? Does it count
in-core cache reads as bandwidth?
Something I do miss, in comparison to iostat and
friends is the occasional new "header" line, telling
me what each column means. Can zpool be extended to
report this every n lines (with default to 0 being
the current behaviour) ?
Darren
_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
