I thought I understood fractional reserve 🙂 until I came across this case, which seems to be an anomaly to me. Is anyone able to explain how the disk space is being used in this example?
Consider two volumes: volproddata03 and volproddiag.
They are both the same size:
% ssh filer df -rk | egrep '(proddiag|proddata03)'
Filesystem total used avail reserved
/vol/volproddata03/ 5033168KB 3357368KB 1675800KB 1258728KB /vol/volproddata03/
snap reserve 1258288KB 772800KB 485488KB 0KB /vol/volproddata03/..
/vol/volproddiag/ 5033168KB 4193344KB 839824KB 2094868KB /vol/volproddiag/
snap reserve 1258288KB 149024KB 1109264KB 0KB /vol/volproddiag/..
but volproddiag is using much more space than volproddata03, despite the fact that the snapshots are smaller:
% ssh filer snap list volproddata03
Volume volproddata03
working...
%/used %/total date name
---------- ---------- ------------ --------
23% (23%) 6% ( 6%) Mar 05 00:55 snproddata03.1
38% (23%) 12% ( 6%) Mar 04 00:55 snproddata03.2
%ssh filer snap list volproddiag
Volume volproddiag
working...
%/used %/total date name
---------- ---------- ------------ --------
3% ( 3%) 1% ( 1%) Mar 05 00:55 snproddiag.1
7% ( 4%) 2% ( 1%) Mar 04 00:55 snproddiag.2
The fractional reserve for both volumes is set to 100% and the guarantee is set to 'volume' in both cases.
Both volumes also contain one LUN each, both the same size, and both LUNs are space reserved.
%ssh filer lun show /vol/volproddiag/volproddiaglun
/vol/volproddiag/volproddiaglun 2.0g (2142240768) (r/w, online, mapped)
%ssh filer lun show /vol/volproddata03/volproddata03lun
/vol/volproddata03/volproddata03lun 2.0g (2142240768) (r/w, online, mapped)
I cannot see any reason why the disk usage for volproddiag (83%) should be higher than volproddata03 (67%)
%ssh filer df -k | egrep '(proddiag|proddata03)'
Filesystem total used avail capacity Mounted on
/vol/volproddata03/ 5033168KB 3357516KB 1675652KB 67% /vol/volproddata03/
snap reserve 1258288KB 772812KB 485476KB 61% /vol/volproddata03/..
/vol/volproddiag/ 5033168KB 4193344KB 839824KB 83% /vol/volproddiag/
snap reserve 1258288KB 149040KB 1109248KB 12% /vol/volproddiag/..
The significant numbers seem to be the 'reserved' column in the 'df -rk' output. However, I cannot see the reason why the reserved space should be so much greater for the volproddiag volume. Is anyone able to explain why?
