Result too large -> no header -> RO hammerfs works

diego cad elmonito at fibertel.com.ar
Thu May 5 22:18:51 PDT 2011


Hi dragons,

i've been around HAMMER for linux past days, after a couple of typos it 
worked nice for Ubuntu running 2.6.37-39. even since the driver had no 
updates since 2009 (?)

1. tut 2 DF running virtualized? i use hless vbox lately with success.

2. still on that 'no header' dmesg sysout on Linux for the relative to \

dragonfoo# mount_hammer: mount on /tmp/fs: Result too large  ->
followup ->    /4d5c172f$0$927$415eb37d at crater_reader.dragonflybsd.org
3. could more ram result into a larger REDO tables ending in succes?
   it is likely that the 'Result too large' for RW mount comes from 
available ram to allocate the REDO tables? then what does that 'no 
header' means (dfbsd and linux) but RO data is untouched.
basically i have 3gig ram and 160g hdd. REDO switches almost 
automatically since last time the box was processing several i mean 
several simultaneous processes (like 40 concurrent RW + 4 makes) then
the REDO as for reordering the last data flushes becomes without enough 
dynamic allocation? it's more likely that it does not relate to HDD 
space as originally thought even if the drive is 25% free capacity under 
that load circumstances.

thx
Diego




More information about the Bugs mailing list