snmp best practices for VNX

Depends on which alerts you want

For events that are file specific you can only get them by configuring the file SNMP – the block SNMP wouldn’t know about them

The other way round – the file side does get a decent amount of the events that happen on the block side

The only snmp traps I am getting right now show up like this:

celWebGUI: Trap message will be sent in the event of a Web GUI notification request. EventId=138

I don’t have any documentation on what the event means.  There is a specific event on the block side in regards to initiators losing connectivity to the storage array, but I’m not sure if that alert would be triggered through the file side.  I have removed the block notification template from the distributed monitors within Unisphere.  Will that force all traps through the file side?

Sure – for file events just look into the VNX Events manual available from support.emc.com

There is also a command in the manual to dump a list of all possible events

Rainer

Is this the same as the "configuring events and notfications" manual?  If so, the event IDs only correlate to the facility, not the error message.

I dont think so. I think through the file side you would only get serious health alerts

I think there is a default template on the block side installed by customer service that is needed for the callhome functionality.

I suggest to work directly with your local service engineer to verify what should be there in your case

Some events have been configured on the file side, but i don't know what the event IDs mean.  Is there a way to correlate the event ID to a specific message?

AFAIK the file side doesnt have any SNMP traps setup by default – only what you specifically configured via nas_event

That's been the problem.  No one really seems to know.  i've been told by EMC support that I shouldn't have both set up and that there's no real way to tell what i'll get from the file side.  This was all set up prior to me starting at this company.

Added bonus for the Unix savvy that get the command to work ☺

Except for dynamox – its too easy for him ☺

$ export NAS_DB=/nas ; /nas/bin/nas_event -l -c -i \ | awk '/^[ ]*[0-9]+/{print

$1}' | xargs -n1 -i bash -c \ "export COMP={} ; /nas/bin/nas_event -l -c {} -i |

awk '/^[ ]*[0-9]+/{print \$1}'\ | xargs -n1 -i /nas/bin/nas_event -l -c \$COMP -f

\{\} -id" | fgrep -v 'DEBUG(7)'

Hint: it used to look nice on the command line with continuation characters but then the formatting process put in different line breaks

Yes – thats the one

It mentions the config files used for the file events – both the default files and the user config

You are correct – the manuals just lists the facilities

If you want to know the event number for a facility just see “man nas_event”

For example this command prints out all possible events for the facility UFS

nas_event -list -component DART -facility UFS

Nas_event can also tell you what events are currently configured for snmp or mail or …

In the manual under “List all events on the control station” there is also a (longer) command to dump all possible events.

I typically just redirect it to a file and manage it via EXcel

Thank you for this information.  I just received a trap and an e-mail alert.  The funny thing is that the e-mail message was "the array is alive".  However the trap came through as "device stlpvnx01-cs0 experienced the following issue: celMasterCtlFault: Trap message will be sent in the event of Master Control serious fault. unexpected daemon exit or control station heartbeat missing. EventId=138".  I'm just curious why the message is not the same.

We are also searching for an solution how to set up Event Management on VNX. Even after reading the document "Configuring Events and Notifications on VNX for File" some questions rose up:

• It seems that all events from the Block backend are forwarded to the the Control Station with the Facility ID 138
• The document doesn't tell any deeper details on this
• Can I create additional Templates on the Block Event Manager (which still appears as the known Navisphere Event Manager)?
• Does an additional template affect the standard "Call Home" template?
• Is it possible to preselect certain events which I want to get informed without impact to the Call Home functionality?

Where can I get more information on this?

Hi mgrammer,

We are also facing the same issue like getting SNMP Trap "Trap message will be sent in the event of Master Control serious faultunexpected daemon exit or control station heartbeat missing."

[1.3.6.1.4.1.1139.2.1.1.4 = Mar 7 13:24:47 2013 PERFSTATS:4:2 nas_alerterd: Raising event for policy: default_nfs_avgtime:server_2:nfs.filesystem.FS-QSilver.client.dhcp-10-16-xx-xx.op.v3Lookup.avgtime is >= 1000000 for 1 minutes. The last sample value was 1548718 ]

But In device i didn't find any logs related to this alert.

Please provide what is the next step to resolve this alert.

Thank you in advance

vajrala