How do I disable individual SMI-S alerts for VMAX arrays?

Maybe you are saying the alert source is set up automatically after you discover an array and it is indeed coming from unisphere.  I don't see the device level alerts in unisphere but I never disabled there either, I will look more there.

Thanks.

Are you sure it's unisphere and not the SMI-S server sending them?  Prosphere doesn't need to talk to unisphere it talks directly to an SMI-S server.

Hi Tazal

The message is reported directly by the SMAS or Unisphere for VMAX SOAP API used by ProSphere alert collection framework. Disable the alert in SMAS or Unisphere for VMAX and in turn the alert should stop appearing in ProSphere.

Opening an SR with ProSphere support if the alert is disabled but still appears in ProPShere

Regards

Conor 

Hi Tazal,

The alert comes from SMC or UniSphere for Vmax (through SMI-S Provider) to ProSphere as the alert is disabled in SMC or UniSphere for Vmax is still appearing in the ProSphere UI open an SR with Support to review why its flooding the alerts view.

Within the help menu of the ProSphere UI goto Home > Operations > General FAQs > What are the different external alerts shown in ProSphere? which reports thin device usage as an alert sent from SMC or UniSphere for Vmax (through SMI-S Provider) to ProSphere.

Regards

Conor

In unisphere going to selecting the array and then choosing administration in the lower right allows me to view what appear to be the alert policies.  The alert for thin device usage is disabled, which appears to be the default, and I do not see the alerts in the unisphere console.

There's either a very simple bug that should always happen or you're misunderstanding how the alerts work.

Can anyone using prosphere (in a real environment) confirm whether or not they get "Thin Devices is now XX percent used" alerts in propshere 2.0?

Thanks.

I see that, thanks for helping me find the proper help section.  Hopefully they will update it to show unisphere as we don't even have SMC installed.

When I read that it tells me that SMI-S is sending it, not SMC/unisphere, but I may be misinterpreting it.

I'll be honest, opening tickets on something like this is painful, which is why I came here first.

Do you know I would change the thresholds for that alert?

I'm including these to assist others who may run in to this in the future.  In the same section of the same doc I found this.  I will try to find the process/script mentioned in the SMI-S documentation and see if that helps me figure out how to eliminate these alerts.

Downgrading to ProSphere 1.0, then upgrading to ProSphere 1.5

When you deploy ProSphere 1.5, ProSphere subscribes to SMC events through the EMC

SMI-S Provider, which converts the SMC events to indications. Even if you remove

ProSphere 1.5, the subscription to the SMI-S Provider indications for SMC events

persists. Suppose you deploy ProSphere 1.0, which does not support consolidation of

SMC alerts, on the same ESX or ESXi cluster, and later you upgrade to ProSphere 1.5.

ProSphere 1.5 now displays all SMC alerts received before the downgrade to ProSphere

1.0, along with the new ones received after the upgrade to ProSphere 1.5.

Until you unsubscribe from SMC events, the SMI-S Provider continues to send indications

to ProSphere. To unsubscribe from SMC events in the SMI-S Provider, you need to

execute an SMI-S script that removes the ProSphere 1.5 subscription. SMI-S

documentation provides information on how to unsubscribe from SMC events.

Hello

Can you please let me know if you are able to unsubscribe "Thin Device Usage alert" in SMI-S ? If yes let me know the document name. I am handling a similar issue.

Thanks in Advance

I was forwarded a document titled "Monitoring EMC Symmetrix Using the Solutions Enabler Event Daemon"
document ID 300-010-522.  It's dated January of 2009 but has information in it about tuning symevent configuration.

I have implemented some test changes and will report back on the results either way.

A couple of interesting things from that document:

You can see a list of event codes like this:

stordaemon action storevntd -cmd list -events

Some of the interesting event codes related to thin provisioning are:

   1111 — Save or data device pool is full.

   1206 — Pool state has changed to [Not Present | Unknown | Online

   | Write Disabled | Offline| Failed].

    1207 — Pool configuration has changed.

    1208 — Pool utilization is now %u percent.

    1212 — Thin device is now %u percent allocated.

    1213 — Thin device is now %u percent used.

    1216 — Pool device state has changed

They gave an example for setting different thresholds for event 1212 and 1213:

Thin device monitoring

             Thin devices can be monitored by the percent that the devices are

             allocated or by the percent used (written). Thresholds can be set to log

             events of varying severity at different percentage levels. The following is

             an example of an entry in the daemon_options file for these events:

storevntd:log_symmetrix_events=sid=000190102000,1212,thresh_critical=80,

                                                thresh_major=65,thresh_warn=50 ; \

sid=000190102000,1213,thresh_critical=80,thresh_major=65, thresh_warn=50 ;

There is also the ability to simple ignore whole event codes.

I have changed some thresholds for one of my arrays to see if the storevnt configuration changes affect the alerts SMI-S sends to prosphere.  If that works I think we will be able to disable the unwanted alerts from being sent to prosphere, if not I will have to go back to support and request an SMI-S person talk to a "regular" solutions enabler person and figure out how to make similar changes to the SMI-S alerting that generic solutions enabler can seemingly make.

FYI: so far I have tested changing the thresholds on one array and completely disabling them on another yet I appear to still be receiving the same types of alerts at the same default thresholds.  It still appears the SMI-S provider is sending events based on a different set of parameters than storevntd.  I am continuing to work with support.

Hi,

I have emailed you a workaround provided by ProSphere engineering to prevent ProSphere from picking up the thin device usage alert by changing the indication_mgr.conf file . This will prevent ProSphere from reporting the alert in the UI. 

Disabling specific set of SMC/UniSPhere alert type is not supported by ProSphere as it pulls them from source.

Thin device usage alerts are received with an EventName:"EMC-SMC-ALERT" from the SMI-S agent.

From the SMI-S agent using the testsmiprovider tool from the options in the Indications menu (ind command) the menu options only allows to block all subscriptions for the destination and it’s not able to filter down to particular event within the subscriptions . It’s possible to block all these but this would block many other SMC/Unisphere alerts from reaching ProSphere.

Regards

Conor

Just to confirm tazal's question, 'yes' Thin Device Usage Alerts still appear in ProSphere 2.0:

Conor,

Could you either post that workaround to this thread or forward it to me? I have the same issue.

Thanks,

Aaron

Hi Aaron,

I have sent you a mail with the file/steps to disable the above alert.

If anyone less requires the workaround please open an SR and reference this forum post.

Thanks

Conor