Sachin Punadikar has uploaded this change for review.
RECOVERY: Do not call grace during shutdown
Call to "rados_grace_join" during shutdown in function
"rados_cluster_shutdown" leads to increamenting current epoch. When the
same Ganesha instance (identified with nodeid) gets started (either on
same node or somewhere else), then current epoch is increased again.
This double increament in epoch leads to not finding recovery object, at
restart of that Ganesha instance.
This happens only for graceful shutdown of Ganesha.
Change-Id: I8e13b117400913c9be9ed894d14216e4dd7b1cdd
Signed-off-by: Sachin Punadikar <sachin.punadikar@ibm.com>
---
M src/SAL/recovery/recovery_rados_cluster.c
1 file changed, 14 insertions(+), 3 deletions(-)
git pull ssh://review.gerrithub.io:29418/ffilz/nfs-ganesha refs/changes/12/1236112/1
To view, visit change 1236112. To unsubscribe, or for help writing mail filters, visit settings.