| Message ID | 1231524865.24591.3.camel@hydrogen.msp.redhat.com (mailing list archive) |
|---|---|
| State | Accepted, archived |
| Delegated to: | Alasdair Kergon |
| Headers | show |
Good that there's no penalty for single node mirroring. ACK Heinz Am Freitag, den 09.01.2009, 12:14 -0600 schrieb Jonathan Brassow: > This patch is the same patch as posted in a previous e-mail. I am > reposting so that it gets into Patchwork. The patch for "cluster-aware > logging module" depends on this. > > brassow > > The logging API needs an extra function to make cluster mirroring > possible. This new function allows us to check whether a mirror > region is being recovered on another machine in the cluster. This > helps us prevent simultaneous recovery I/O and process I/O to the > same locations on disk. > > Cluster-aware log modules will implement this function. Single > machine log modules will not. So, there is no performance > penalty for single machine mirrors. > > Signed-off-by: Jonathan Brassow <jbrassow@redhat.com> > > Index: linux-2.6/drivers/md/dm-raid1.c > =================================================================== > --- linux-2.6.orig/drivers/md/dm-raid1.c > +++ linux-2.6/drivers/md/dm-raid1.c > @@ -586,6 +586,9 @@ static void do_writes(struct mirror_set > int state; > struct bio *bio; > struct bio_list sync, nosync, recover, *this_list = NULL; > + struct bio_list requeue; > + struct dm_dirty_log *log = dm_rh_dirty_log(ms->rh); > + region_t region; > > if (!writes->head) > return; > @@ -596,10 +599,18 @@ static void do_writes(struct mirror_set > bio_list_init(&sync); > bio_list_init(&nosync); > bio_list_init(&recover); > + bio_list_init(&requeue); > > while ((bio = bio_list_pop(writes))) { > - state = dm_rh_get_state(ms->rh, > - dm_rh_bio_to_region(ms->rh, bio), 1); > + region = dm_rh_bio_to_region(ms->rh, bio); > + > + if (log->type->is_remote_recovering && > + log->type->is_remote_recovering(log, region)) { > + bio_list_add(&requeue, bio); > + continue; > + } > + > + state = dm_rh_get_state(ms->rh, region, 1); > switch (state) { > case DM_RH_CLEAN: > case DM_RH_DIRTY: > @@ -619,6 +630,16 @@ static void do_writes(struct mirror_set > } > > /* > + * Add bios that are delayed due to remote recovery > + * back on to the write queue > + */ > + if (unlikely(requeue.head)) { > + spin_lock_irq(&ms->lock); > + bio_list_merge(&ms->writes, &requeue); > + spin_unlock_irq(&ms->lock); > + } > + > + /* > * Increment the pending counts for any regions that will > * be written to (writes to recover regions are going to > * be delayed). > Index: linux-2.6/include/linux/dm-dirty-log.h > =================================================================== > --- linux-2.6.orig/include/linux/dm-dirty-log.h > +++ linux-2.6/include/linux/dm-dirty-log.h > @@ -113,6 +113,16 @@ struct dm_dirty_log_type { > */ > int (*status)(struct dm_dirty_log *log, status_type_t status_type, > char *result, unsigned maxlen); > + > + /* > + * is_remote_recovering is necessary for cluster mirroring. It provides > + * a way to detect recovery on another node, so we aren't writing > + * concurrently. This function is likely to block (when a cluster log > + * is used). > + * > + * Returns: 0, 1 > + */ > + int (*is_remote_recovering)(struct dm_dirty_log *log, region_t region); > }; > > int dm_dirty_log_type_register(struct dm_dirty_log_type *type); > > > -- > dm-devel mailing list > dm-devel@redhat.com > https://www.redhat.com/mailman/listinfo/dm-devel -- dm-devel mailing list dm-devel@redhat.com https://www.redhat.com/mailman/listinfo/dm-devel
Index: linux-2.6/drivers/md/dm-raid1.c =================================================================== --- linux-2.6.orig/drivers/md/dm-raid1.c +++ linux-2.6/drivers/md/dm-raid1.c @@ -586,6 +586,9 @@ static void do_writes(struct mirror_set int state; struct bio *bio; struct bio_list sync, nosync, recover, *this_list = NULL; + struct bio_list requeue; + struct dm_dirty_log *log = dm_rh_dirty_log(ms->rh); + region_t region; if (!writes->head) return; @@ -596,10 +599,18 @@ static void do_writes(struct mirror_set bio_list_init(&sync); bio_list_init(&nosync); bio_list_init(&recover); + bio_list_init(&requeue); while ((bio = bio_list_pop(writes))) { - state = dm_rh_get_state(ms->rh, - dm_rh_bio_to_region(ms->rh, bio), 1); + region = dm_rh_bio_to_region(ms->rh, bio); + + if (log->type->is_remote_recovering && + log->type->is_remote_recovering(log, region)) { + bio_list_add(&requeue, bio); + continue; + } + + state = dm_rh_get_state(ms->rh, region, 1); switch (state) { case DM_RH_CLEAN: case DM_RH_DIRTY: @@ -619,6 +630,16 @@ static void do_writes(struct mirror_set } /* + * Add bios that are delayed due to remote recovery + * back on to the write queue + */ + if (unlikely(requeue.head)) { + spin_lock_irq(&ms->lock); + bio_list_merge(&ms->writes, &requeue); + spin_unlock_irq(&ms->lock); + } + + /* * Increment the pending counts for any regions that will * be written to (writes to recover regions are going to * be delayed). Index: linux-2.6/include/linux/dm-dirty-log.h =================================================================== --- linux-2.6.orig/include/linux/dm-dirty-log.h +++ linux-2.6/include/linux/dm-dirty-log.h @@ -113,6 +113,16 @@ struct dm_dirty_log_type { */ int (*status)(struct dm_dirty_log *log, status_type_t status_type, char *result, unsigned maxlen); + + /* + * is_remote_recovering is necessary for cluster mirroring. It provides + * a way to detect recovery on another node, so we aren't writing + * concurrently. This function is likely to block (when a cluster log + * is used). + * + * Returns: 0, 1 + */ + int (*is_remote_recovering)(struct dm_dirty_log *log, region_t region); }; int dm_dirty_log_type_register(struct dm_dirty_log_type *type);
This patch is the same patch as posted in a previous e-mail. I am reposting so that it gets into Patchwork. The patch for "cluster-aware logging module" depends on this. brassow The logging API needs an extra function to make cluster mirroring possible. This new function allows us to check whether a mirror region is being recovered on another machine in the cluster. This helps us prevent simultaneous recovery I/O and process I/O to the same locations on disk. Cluster-aware log modules will implement this function. Single machine log modules will not. So, there is no performance penalty for single machine mirrors. Signed-off-by: Jonathan Brassow <jbrassow@redhat.com> -- dm-devel mailing list dm-devel@redhat.com https://www.redhat.com/mailman/listinfo/dm-devel