Active IQ Unified Manager Discussions
Active IQ Unified Manager Discussions
Hi Gurus
I found something I don't understand about how DFM define which events will be triggered and which ones are not. for example, I am testing the events about snapmirror. totally there are following eventtype:
snapmirror-abort:completed | Normal | sm.abort |
snapmirror-abort:failed | Error | sm.abort |
snapmirror-break:completed | Normal | sm.break |
snapmirror-break:failed | Error | sm.break |
snapmirror-delete:aborted | Warning | sm.delete |
snapmirror-delete:completed | Normal | sm.delete |
snapmirror-delete:failed | Error | sm.delete |
snapmirror-initialize:aborted | Warning | sm.initialize |
snapmirror-initialize:completed | Normal | sm.initialize |
snapmirror-initialize:failed | Error | sm.initialize |
snapmirror-quiesce:aborted | Warning | sm.quiesce |
snapmirror-quiesce:completed | Normal | sm.quiesce |
snapmirror-quiesce:failed | Error | sm.quiesce |
snapmirror-resume:completed | Normal | sm.resume |
snapmirror-resume:failed | Error | sm.resume |
snapmirror-resync:aborted | Warning | sm.resync |
snapmirror-resync:completed | Normal | sm.resync |
snapmirror-resync:failed | Error | sm.resync |
snapmirror-update:aborted | Warning | sm.update |
snapmirror-update:completed | Normal | sm.update |
snapmirror-update:failed | Error | sm.update |
snapmirror:date-ok | Normal | sm.lag |
snapmirror:deleted | Information sm.lag | |
snapmirror:discovered | Information sm.relationship | |
snapmirror:modified | Information sm.relationship | |
snapmirror:nearly-out-of-date | Warning | sm.lag |
snapmirror:not-scheduled | Normal | sm.state |
snapmirror:not-working | Error | sm.state |
snapmirror:off | Normal | sm.state |
snapmirror:out-of-date | Error | sm.lag |
snapmirror:possible-problem | Warning | sm.state |
snapmirror:undeleted | Information sm.relationship | |
snapmirror:unknown-state | Warning | sm.state |
snapmirror:working | Normal | sm.state |
but when I tried to initialize a snapmirror relation, when initialize,abort it, initialize it again, abort it, resync it . it should generate many events like snapmirror-initialize:completed, snapmirror-resync:completed. but in my case, only snapmirror:discovered is triggered. is it expected? your input are appreciated.
Thanks and Best Regards!
TC
Hi TC,
When you say I initialized, abort etc where did you do it from ? Is it from Protection Manger or Disaster Recovery Manager ? If its done from the PM/DRM you will see all these events else you will only see below to events.
snapmirror:date-ok | Normal | sm.lag |
snapmirror:nearly-out-of-date | Warning | sm.lag |
Same is the case when you do it from filer view/SystemManager/Filer cli.
You can also check the following thread on the same topic.
https://communities.netapp.com/message/93280#93280
Regards
adai
Hi Adai
All operations were done via CLI. but both source and destination are under same DFM and NMC could see the snapmirror relation but as external relation.
if customer want to see snapmirror alert either CLI operation or PM, what should we configure? import snapmirror relation to a dataset? or anything else?
what's the document to describe which event will show-up or which ones will not? I see it's no related with severity.
BR
TC
The reason we dont generate the events you are expecting is, these relationships are external. We generate all the events only when dfm manages these releationships and not when it only monitors.
Yes you will have to import them into a dataset in order for dfm to manage it.
Regards
adai
I tried to initialize a snapmirror, before initialization is completed, I abort it. I assume the followin event will be triggered:
snapmirror-initialize:aborted | Warning | sm.initialize |
this is warning event.
but it did not happen.
Hi TC
Thats exactly what I said, if you do these in filer cli and not with dfdrm or dfpm cli you will not receive any snapmirror events other than lag error and warning events.
Regards
adai
Hi Adai
Thanks for your kindly reply.
I just wonder
snapmirror-initialize:aborted | Warning | sm.initialize |
this event is warning event. even I abort the initialize by CLI, DFM should receive event? right?
Do you think if there is anything I missed? I am still confused.
1) external snapmirror relation, abort intialize by CLI, will DFM receive event (warning event)?
2) import snapmirror relation to dataset, but still abort it by CLI, will DFM receive event?
3) import snapmirror relation to dataset, operation by PM, will DFM receive event? (any level?)
is there a deep KB or document about that?
BR
TC
HI TC,
I dont know what is not clear. Can you pls read this bug 206891 ? DFM is monitors for SnapMirror related things once every 30mins, so it can only report on the lag times of a relationship.
In order for dfm to generate these events either dfm should get traps or the action itself should be initiated from dfm so that it can generate the events. Hope this helps and clarifies.
1) external snapmirror relation, abort intialize by CLI, will DFM receive event (warning event)?
No, because dfm doesn't get a trap and it only monitors once every 30mins.
2) import snapmirror relation to dataset, but still abort it by CLI, will DFM receive event?
No. Same as above.
3) import snapmirror relation to dataset, operation by PM, will DFM receive event? (any level?)
Yes. Since the operation is initiated from dfm/PM it will receive the response and trigger the event.
Regards
adai
Adai: Thank you very much!
Hi Adai
Sorry for further question. for the below events:
snapmirror:date-ok | Normal | sm.lag |
snapmirror:nearly-out-of-date | Warning | sm.lag |
if snapmirror relation (update interval) are defined in filer, How DFM receive the lag error? how DFM determin a snapmirror relation lag time? (for example, one snapmirror update interval is 6 hours, and another one is 12 hours).
I know PM could define a dataset warning lag time. but a snapmirror relation is not imported to a dataset. DFM could also receive the lag warning. right?
Hi Chao,
The default threshold for snapmirror:nearly-out-of-date is 1.5days or 36hours and snapmirror:out-of-date is 2days or 48 hours.
If you want to get the alerts based on your scheduling then you will have to import them to dataset and define new lag times based on scheduling for each dataset.
Regards
adai
I see. thanks a lot. Sounds we could not avoid PM.