Lines Matching defs:node
49 function being executed on the remote node. The connecting send/receive
448 struct rb_node *node = tree->rb_node;
452 while (node) {
453 r = rb_entry(node, struct dlm_rsb, res_hashnode);
456 node = node->rb_left;
458 node = node->rb_right;
519 * (we are the dir node for the res, but are not using the res right now,
520 * but some other node is)
573 * If someone sends us a request, we are the dir node, and we do
620 /* our rsb was not master, and another node (not the dir node)
700 node immediately */
763 /* our rsb is not master, and another node has sent us a
853 /* our rsb is not master, and another node (not the dir node)
882 * We're the dir node for this res and another node wants to know the
884 * called from receive_lookup(); master lookups when the local node is
885 * the dir node are done by find_rsb().
887 * normal operation, we are the dir node for a resource
897 * remote node sends back the rsb names it is master of and we are dir of
899 * we either create new rsb setting remote node as master, or find existing
900 * rsb and set master to be the remote node.
1024 /* this can happen when the master sends remove, the dir node
1435 "node %d", lkb->lkb_id, (long long)us,
1448 a reply from a remote node */
1675 for the master node to send us a dir remove for
1697 dir node to remove the dir record. */
1729 * important to keep us (the master node) from being out of sync
1730 * with the remote dir node for very long.
1907 log_debug(ls, "timeout cancel %x node %d %s",
2143 be sent to the requesting node in addition to granting the lock if the
2144 lkb belongs to a remote node. */
2533 log_print("WARN: pending demoted %x node %d %s",
2551 log_print("WARN: pending deadlock %x node %d %s",
2739 there are cases where we become the dir node after we've
3073 * These are called on the master node for the given lock and
3243 /* receive_request() calls do_request() on remote node */
3262 /* receive_convert() calls do_convert() on remote node */
3281 /* receive_unlock() calls do_unlock() on remote node */
3300 /* receive_cancel() calls do_cancel() on remote node */
4106 /* The dir node is the authority on whether we are the master
4109 send a remove message to the dir node at the same time that the dir
4110 node sends us a request for the rsb. */
4155 delayed in being sent/arriving/being processed on the dir node.
4156 Another node would repeatedly lookup up the master, and the dir
4157 node would continue returning our nodeid until our send_remove
4425 message. This is an expected race between the dir node sending a
4426 request to the master node at the same time as the master node sends
4427 a remove to the dir node. The resolution to that race is for the
4428 dir node to ignore the remove message, and the master node to
4429 recreate the master rsb when it gets a request from the dir node for
4520 /* Optimization: the dir node was also the master, so it took our
4806 /* ms->m_result is the value returned by dlm_master_lookup on dir node
4819 /* We sometimes receive a request from the dir node for this
4820 rsb before we've received the dir node's loookup_reply for it.
4821 The request from the dir node implies we're the master, so we set
4840 /* the remote node doesn't believe it's the dir node */
4885 /* messages sent to a master node */
4904 /* messages sent from a master node (replies to above) */
4922 /* messages sent from a master node (only two types of async msg) */
4934 /* messages sent to a dir node */
4944 /* messages sent from a dir node (remove has no reply) */
5111 /* A waiting lkb needs recovery if the master node has failed, or
5112 the master node is changing (only when no directory is used) */
5128 dead node. Requests and up-conversions we flag to be resent after
5261 master or dir-node for r. Processing the lkb may result in it being placed
5423 because a node holding EX/PW failed */
5453 case of a single node removed */
5658 have to require it. Recovery of masters on one node can overlap
5659 recovery of locks on another node, so one node can send us MSTCPY