Ocfs2 devel - May 2011 - [PATCH] ocfs2/dlm: Target node death during resource migration leads to thread spin

During resource migration, if the target node were to die, the thread doing
the migration spins until the target node is not removed from the domain map.
This patch slows the spin by making the thread wait for the recovery to kick in.

Signed-off-by: Sunil Mushran <sunil.mushran at oracle.com>
---
 fs/ocfs2/dlm/dlmmaster.c |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)

diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c
index 4499d86..4a780a3 100644
--- a/fs/ocfs2/dlm/dlmmaster.c
+++ b/fs/ocfs2/dlm/dlmmaster.c
@@ -2522,6 +2522,9 @@ fail:
 		res->state &= ~DLM_LOCK_RES_MIGRATING;
 		wake = 1;
 		spin_unlock(&res->spinlock);
+		if (dlm_is_host_down(ret))
+			dlm_wait_for_node_death(dlm, target,
+						DLM_NODE_DEATH_WAIT_MAX);
 		goto leave;
 	}
 
-- 
1.7.4.1

Joel,

FYI

The list of pending patches from my end are available here:
http://oss.oracle.com/~smushran/ocfs2-pending-patches/fs-may04/

Thanks
Sunil

On 05/04/2011 10:34 AM, Sunil Mushran wrote:
> During resource migration, if the target node were to die, the thread doing
> the migration spins until the target node is not removed from the domain
map.
> This patch slows the spin by making the thread wait for the recovery to
kick in.
>
> Signed-off-by: Sunil Mushran<sunil.mushran at oracle.com>
> ---
>   fs/ocfs2/dlm/dlmmaster.c |    3 +++
>   1 files changed, 3 insertions(+), 0 deletions(-)
>
> diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c
> index 4499d86..4a780a3 100644
> --- a/fs/ocfs2/dlm/dlmmaster.c
> +++ b/fs/ocfs2/dlm/dlmmaster.c
> @@ -2522,6 +2522,9 @@ fail:
>   		res->state&= ~DLM_LOCK_RES_MIGRATING;
>   		wake = 1;
>   		spin_unlock(&res->spinlock);
> +		if (dlm_is_host_down(ret))
> +			dlm_wait_for_node_death(dlm, target,
> +						DLM_NODE_DEATH_WAIT_MAX);
>   		goto leave;
>   	}

On Wed, May 04, 2011 at 10:34:49AM -0700, Sunil Mushran
wrote:
> During resource migration, if the target node were to die, the thread doing
> the migration spins until the target node is not removed from the domain
map.
> This patch slows the spin by making the thread wait for the recovery to
kick in.
> 
> Signed-off-by: Sunil Mushran <sunil.mushran at oracle.com>
Signed-off-by: Mark Fasheh <mfasheh at suse.com>

--
Mark Fasheh