[Lustre-devel] [RFC] "lctl readonly" modification proposal

Alexander Zarochentsev Alexander.Zarochentsev at Sun.COM
Thu Aug 21 02:24:08 PDT 2008


On 20 August 2008 23:42:39 Mikhail Pershin wrote:
> On Wed, 20 Aug 2008 21:39:13 +0400, Peter Braam <Peter.Braam at sun.com>
>
> wrote:
> > If I remember correctly the flush is only there to try to reduce
> > rollback.
>
> Also to be sure that all before barrier is on disk, many test cases
> about recovery rely on the fact that operations before barrier are on
> disk. This is important with HARD failure case especially, without
> sync before barrier the operations before barrier may be lost and
> affect the tests. I agree that syncless variant of barrier is needed
> but we also need barrier with sync depending on test case.

You don't need lctl readonly to do a sync, the replay_barrier function 
from test-framework.sh adds a sync explicitly:

replay_barrier() {
    local facet=$1
    do_facet $facet sync
    df $MOUNT
    local svc=${facet}_svc
    do_facet $facet $LCTL --device %${!svc} readonly
    do_facet $facet $LCTL --device %${!svc} notransno
    do_facet $facet $LCTL mark "$facet REPLAY BARRIER on ${!svc}"
    $LCTL mark "local REPLAY BARRIER on ${!svc}"
}

WBR,
-- 
Alexander "Zam" Zarochentsev
Staff Engineer
Lustre Group, Sun Microsystems



More information about the lustre-devel mailing list