diff options
| author |  Jeffrey Hutzelman <jhutz@cmu.edu> | 2013-02-07 12:23:50 -0500 |
|---|---|---|
| committer | Jeffrey Hutzelman <jhutz@cmu.edu> | 2013-02-14 19:57:31 -0500 |
| commit | f61ee33be0dde91bc33772fabe3300d98527b29e (patch) | |
| tree | 671090ddfc7776470e25544c5002723a0dd68a53 | |
| parent | e2bfb6322ce9f4323b83469dc06dd1ce1b0f4cf6 (diff) | |
Avoid hello storms when a bdump fails
When a braindump fails, leave the other server's status at SERV_STARTING
instead of downgrading to SERV_DEAD, and schedule the next hello to that
server at the regular timeout instead of immediately. Otherwise, failed
bdumps are retried repeatedly with no delay, keeping the server busy, the
logs full, and the network congested.
This fixes #89
| -rw-r--r-- | server/bdump.c | 5 |
1 files changed, 3 insertions, 2 deletions
diff --git a/server/bdump.c b/server/bdump.c index c790d8f..5e7b604 100644 --- a/server/bdump.c +++ b/server/bdump.c @@ -966,9 +966,10 @@ cleanup(Server *server) zdbug((LOG_DEBUG, "bdump cleanup")); if (server != limbo_server) { - server->state = SERV_DEAD; + if (server->state != SERV_STARTING) + server->state = SERV_DEAD; timer_reset(server->timer); - server->timer = timer_set_rel(0L, server_timo, server); + server->timer = timer_set_rel(server->timeout, server_timo, server); } shutdown_file_pointers (); #ifdef _POSIX_VERSION |