diff options
| author | Sage Weil <sage@inktank.com> | 2012-07-10 18:53:34 (GMT) |
|---|---|---|
| committer | Sage Weil <sage@inktank.com> | 2012-07-30 16:29:52 (GMT) |
| commit | a16cb1f70799c851410d9dca0a24122e258df06c (patch) | |
| tree | 0290805d3594d770da3937da7fa68841c1e294ac /include/linux/ceph | |
| parent | cd43045c2de60f40a0aea49bfb252a2eafe58f8c (diff) | |
| download | linux-fsl-qoriq-a16cb1f70799c851410d9dca0a24122e258df06c.tar.xz | |
libceph: fix messenger retry
In ancient times, the messenger could both initiate and accept connections.
An artifact if that was data structures to store/process an incoming
ceph_msg_connect request and send an outgoing ceph_msg_connect_reply.
Sadly, the negotiation code was referencing those structures and ignoring
important information (like the peer's connect_seq) from the correct ones.
Among other things, this fixes tight reconnect loops where the server sends
RETRY_SESSION and we (the client) retries with the same connect_seq as last
time. This bug pretty easily triggered by injecting socket failures on the
MDS and running some fs workload like workunits/direct_io/test_sync_io.
Signed-off-by: Sage Weil <sage@inktank.com>
Diffstat (limited to 'include/linux/ceph')
| -rw-r--r-- | include/linux/ceph/messenger.h | 12 |
1 files changed, 2 insertions, 10 deletions
diff --git a/include/linux/ceph/messenger.h b/include/linux/ceph/messenger.h index 478f814..cfb1bbd 100644 --- a/include/linux/ceph/messenger.h +++ b/include/linux/ceph/messenger.h @@ -172,16 +172,8 @@ struct ceph_connection { /* connection negotiation temps */ char in_banner[CEPH_BANNER_MAX_LEN]; - union { - struct { /* outgoing connection */ - struct ceph_msg_connect out_connect; - struct ceph_msg_connect_reply in_reply; - }; - struct { /* incoming */ - struct ceph_msg_connect in_connect; - struct ceph_msg_connect_reply out_reply; - }; - }; + struct ceph_msg_connect out_connect; + struct ceph_msg_connect_reply in_reply; struct ceph_entity_addr actual_peer_addr; /* message out temps */ |