KernelPanic wrote at 2008-01-23 13:31:

Some problems compiling on a 2.6.23 kernel, im using a debian unstable distro.

CC [M] drivers/block/dst/dcore.o drivers/block/dst/dcore.c: In function ‘dst_sys_init’: drivers/block/dst/dcore.c:1597: error: too many arguments to function ‘kmem_cache_crea

zbr wrote at 2008-01-23 14:31:

That's because I still use 2.6.22 :)

Serge Leschinsky wrote at 2008-02-08 10:14:

I'm sorry, but it's impossible to send you the testing results

----- The following addresses had permanent fatal errors ----- <johnpol@xxx.mipt.ru> (reason: Can't create output)

----- Transcript of session follows ----- procmail: No space left to finish writing "/var/spool/mail/johnpol" 550 5.0.0 <johnpol@xxx.mipt.ru>... Can't create output

zbr wrote at 2008-02-08 10:24:

I kicked admin and fixed this stuff by hands, thanks a lot for pointing!

dmy wrote at 2008-03-11 19:20:

Hi,

Is it possible to use dst in synchronous mode for alg_mirror ?

I mean, a write to the dst-device only returns once all mirrored nodes also performed the write (except perhaps if a node is down).

zbr wrote at 2008-03-11 19:30:

Not yet, barriers are not supported in dst, it is in todo list, but not with a high priority, iirc no other multiple device mappers (except drbd) support barriers.

dmy wrote at 2008-03-12 00:25:

Thanks, I did not get it from the TODO list.

I hope your illness is now part of the past, good devel blogs are not so common ;)

Zbr wrote at 2008-03-12 00:37:

Thanks :) I think I'm about to be ready to start active hacking again, although I have not yet finished watching all 4 seasons of 'House M. D.' :)

DST currently has two opened bugs reported to me which I will fix asap when I get access to hardware to test patches on. I expected it to arrive previous week, but that did not happen, so expectation moved to March 8-15. Per-project todo lists are not published in the main one on top of the page, that is usually only for major projects.