On 12/14/18 9:03 AM, Dongsheng Yang wrote:
On 12/13/2018 06:41 PM, cgxu519 wrote:
On 12/13/18 5:48 PM, Dongsheng Yang wrote:
Introduce a new option abort_on_full, default to false. Then
we can get -ENOSPC when the pool is full, or reaches quota.
---
fs/ceph/super.c | 2 +-
include/linux/ceph/libceph.h | 1 +
net/ceph/ceph_common.c | 10 ++++++++++
3 files changed, 12 insertions(+), 1 deletion(-)
diff --git a/fs/ceph/super.c b/fs/ceph/super.c
index b5ecd6f..f97857a 100644
--- a/fs/ceph/super.c
+++ b/fs/ceph/super.c
@@ -632,6 +632,7 @@ static struct ceph_fs_client
*create_fs_client(struct ceph_mount_options *fsopt,
goto fail;
}
+ opt->flags |= CEPH_OPT_ABORTONFULL;
According to your change log, it should not be set here.
Do you mean "default to false"?
No, the default to false mean the default in libceph is false, please
take
a look at CEPH_OPT_DEFAULT.
But there are two user of libceph, cephfs and krbd. cephfs was
always using abort_on_full, so set this flag in cephfs.
Yeah, you are right. Please ignore that.
Thanks,
