Hi,
I have strange issue when using API over ontap 8 and ontap 9. My setup is one metrocluster ontap 8 and one cluster ontap 9.
Poblem indicate its self, while collecting statistics for Zenoss, and like, broken responses for components like disk, volumes and LUN-s, but not any request fails.
There is situatin that I recive results for example of 10 requests and then 2 are broken...as broken deffered function.
Also components of passive instance of metrocluser work just fine, without any hole, also Spare diska as fine, Cluster overview stats (Cluster)..
Does anyone have experience with with situation like this one.
Not sure but my idea is that API have low privileges calls, like retriving statistics from disk (btw I'm quering all disks, volumes, and LUNs at ones, collection of items for any component), and storage protects itsself from kind of DOS?
perfGet = {
'lun':
dict.fromkeys(
[ds.component for ds in config.datasources],
[
('avg_latency', 'AVERAGE', 'total_ops'),
('avg_read_latency', 'AVERAGE','read_ops'),
('avg_write_latency', 'AVERAGE', 'write_ops'),
('read_data', 'RATE'),
('write_data', 'RATE'),
]
)
}
And this is my query maker:
def _generate_request(self, perfGet):
request = NaElement('perf-object-get-instances')
for perf, instances in perfGet.iteritems():
request.child_add_string('objectname', perf)
reqinst = NaElement('instance-uuids')
for instance, counters in instances.iteritems():
reqinst.child_add_string('instance-uuid', instance)
reqcnts = NaElement('counters')
for counter in counters:
reqcnts.child_add_string('counter', counter[0])
try:
reqcnts.child_add_string('counter', counter[2])
except Exception as ass:
pass
request.child_add(reqcnts)
request.child_add(reqinst)
return request
Where list of components are all luns in system.
Any ides would be useful, becuase after debugging looks like I'm in deadend.
Thanks
Dubravko