Am 28. Mai 2010 03:59 schrieb Tom Limoncelli <[email protected]>:
> --- a/daemons/ganeti-watcher
> +++ b/daemons/ganeti-watcher
> @@ -48,6 +48,7 @@ from ganeti import ssconf
> - utils.EnsureDaemon(constants.RAPI)
> +
> + # Restart RAPI if it isn't responding to queries.
> + # Only kill/restart RAPI once. Otherwise just give up.
> + rapi_restarted = False
> + while True:
> + utils.EnsureDaemon(constants.RAPI)
> + logging.debug("Attempting to talk with RAPI")
> + master_rapi = rapi_client.GanetiRapiClient("localhost",
> + ssl_cert_file=constants.RAPI_CERT_FILE)
This won't work on master HEAD. Please implement proper SSL
certificate verification (see tools/move-instance).
> + try:
> + master_version = master_rapi.GetVersion()
> + except:
Catch specific exceptions.
> + logging.error("Could not open connection to RAPI")
logging.exception
> + if rapi_restarted:
> + break
> + else:
> + logging.debug("RAPI is running but did not speak. Killing
> RAPI")
> + utils.StopDaemon(constants.RAPI)
> + continue
> + if master_version == 2:
Don't hardcode this, move it to a constant at least.
> + break
> + else:
> + logging.fatal("RAPI version said %s, expecting 2" % master_version)
>
> try:
> watcher = Watcher(options, notepad)
> diff --git a/lib/utils.py b/lib/utils.py
> index 2b3e785..8594779 100644
> --- a/lib/utils.py
> +++ b/lib/utils.py
> @@ -2148,6 +2148,19 @@ def EnsureDaemon(name):
> return True
>
>
> +def StopDaemon(name):
> + """Stop a daemon.
> +
> + """
> + result = RunCmd([constants.DAEMON_UTIL, "stop", name])
Why not “restart”?
Michael
