Merge "Add an exception handler to deal with errors that occur when the info in nova is out of sync with reality (as on my currently broken system). Also adds a nova prefix to the logger for now so messages from this module make it into the log file."
This commit is contained in:
commit
68ae03c387
@ -29,7 +29,10 @@ import nova.notifier.rabbit_notifier
|
||||
from ceilometer import rpc
|
||||
|
||||
FLAGS = flags.FLAGS
|
||||
LOG = logging.getLogger(__name__)
|
||||
# FIXME(dhellmann): We need to have the main program set up logging
|
||||
# correctly so messages from modules outside of the nova package
|
||||
# appear in the output.
|
||||
LOG = logging.getLogger('nova.' + __name__)
|
||||
|
||||
|
||||
class InstanceManager(manager.Manager):
|
||||
@ -51,8 +54,9 @@ class ComputeManager(manager.Manager):
|
||||
domain = conn._conn.lookupByName(instance)
|
||||
tree = etree.fromstring(domain.XMLDesc(0))
|
||||
return filter(bool,
|
||||
[target.get('dev') \
|
||||
for target in tree.findall('devices/disk/target')])
|
||||
[target.get('dev')
|
||||
for target in tree.findall('devices/disk/target')
|
||||
])
|
||||
|
||||
@manager.periodic_task
|
||||
def _fetch_diskio(self, context):
|
||||
@ -62,6 +66,12 @@ class ComputeManager(manager.Manager):
|
||||
self.host):
|
||||
# TODO(jd) This does not work see bug#998089
|
||||
# for disk in conn.get_disks(instance.name):
|
||||
for disk in self._get_disks(conn, instance.name):
|
||||
try:
|
||||
disks = self._get_disks(conn, instance.name)
|
||||
except Exception as err:
|
||||
LOG.warning('Ignoring instance %s: %s', instance.name, err)
|
||||
LOG.exception(err)
|
||||
continue
|
||||
for disk in disks:
|
||||
stats = conn.block_stats(instance.name, disk)
|
||||
LOG.info("DISKIO USAGE: %s %s: read-requests=%d read-bytes=%d write-requests=%d write-bytes=%d errors=%d" % (instance, disk, stats[0], stats[1], stats[2], stats[3], stats[4]))
|
||||
|
Loading…
Reference in New Issue
Block a user