Lines Matching full:we
7 /* This quorum hack is only here until we transition to some more rational
17 * So we declare that a node which has given up on connecting to a majority
20 * There are huge opportunities for races here. After we give up on a node's
21 * connection we need to wait long enough to give heartbeat an opportunity
22 * to declare the node as truly dead. We also need to be careful with the
23 * race between when we see a node start heartbeating and when we connect
78 * other nodes in the cluster may consider us dead at that time so we
79 * want to "fence" ourselves so that we don't scribble on the disk
82 * least close some of those gaps. When we have real fencing, this can
112 * if we can't talk to the majority we're hosed */ in o2quo_make_decision()
125 * that case we're hosed if we can't talk to the group that has in o2quo_make_decision()
191 /* as a node comes up we delay the quorum decision until we know the fate of
193 * perpetuated by con_err until hb_down. if we already have a conn, we might
217 /* hb going down releases any holds we might have had due to this node from
239 /* this tells us that we've decided that the node is still heartbeating
240 * even though we've lost it's conn. it must only be called after conn_err
241 * and indicates that we must now make a quorum decision in the future,
242 * though we might be doing so after waiting for holds to drain. Here
243 * we'll be dropping the hold from conn_err. */
258 /* This is analogous to hb_up. as a node's connection comes up we delay the
259 * quorum decision until we see it heartbeating. the hold will be droped in
261 * it's already heartbeating we might be dropping a hold that conn_up got.
285 /* we've decided that we won't ever be connecting to the node again. if it's
286 * still heartbeating we grab a hold that will delay decisions until either the