This post is more than 5 years old

30 Posts

2822

December 23rd, 2009 14:00

Failed Disk and no Alerts in Navisphere?

In Navisphere I get alerts when I lose a drive in my CX700 but not my new CX4960.  No icon changes, no alerts in the alert window. Nothing.  I do get the orange light on the CX4960 unit in the Data Center but Navisphere is oblivious. If I right click on the SP and go to events then I see a Critical icon warning about a disk failure.  It would be nice if my new supper whammy CX4960 let me know there was a bad drive.   Am I missing something?

Thank you.

34 Posts

December 29th, 2009 09:00

This issue is described in the FLARE release notes,

Quoting from the release notes:

"

Faulted unbound drive

If a disk drive is in a faulted state, but not part of a bound LUN, then the drive fault LED will be lit and a log entry will indicate the drive is faulted, but Navisphere Manager will not display the drive as in the faulted state. It will show the enclosure as faulted. If the drive is part of a bound LUN, then Navisphere Manager will show the drive as faulted.

"

I have no idea if this condition would not dial out or email out.

Cheers,

-joji

6 Operator

 • 

2.1K Posts

December 24th, 2009 08:00

Did you buy the Alerting module with the array?

Just kidding! Actually this does not make a lot of sense. I don't have a 960 (until Jan 15th) but I've got several CX4s on the floor and I've never had a problem having alerts show up for any type of failure. What version of FLARE are you running?

6 Operator

 • 

2.1K Posts

December 24th, 2009 08:00

BTW - Welcome to the Forums community! I can see you've been here for a while, but thanks for joining in with your first post. Hopefully we can collectively get to the bottom of this issue with you.

I didn't think to ask the first time around, but did the array dial home? If so, did you mention this to the support tech who responded?

30 Posts

December 29th, 2009 08:00

Thank you for the response.  I apologize for the delay, I took some time off for Christmas.

My FLARE is 4.29.000.5.003.

Far as I can tell, the CLARiiON did not call home.  I have my email address in the CC section of the call home template and did not receive anything.  I do receive the test emails when I initiate them from the template.  Should EMC respond to the tests?  They never have.  So I’m left wondering if EMC gets any email from my CLARiiON.  I verified in Exchange that the test emails to emailalert@emc.com do indeed go out.

About the bad drive, when I go into Navisphere everything looks good.  The alert window is empty and no warning icons on the CX4960.  However, if I drill down into Physical, Bus Enclosure, Disks, I immediately get a purple “R” with a circle next to the bad disk.  This also triggers a red triangle with an “!” in it to be overlaid on the “Disks”, “Bus Enclosure”, “physical” and all the way up to the “Storage Domains”. And it stays lit up like a Christmas tree until I logout and log back in.

The DAE has no RAID groups assigned to it yet and all the drives are unallocated.  Maybe this is why the bad drive was automatically removed. My guess is that when a non-used drive is automatically removed due to failure it does not trigger an alert?  Just an ugly discomforting bright orange light on the drive.  I could test my theory buy carving out a test LUN then manually removing a drive and see if I get an Alert.

Any thoughts?

As you have noticed, I am new here.  So please forgive my ignorance.  I’ve signed myself up for a couple week long CLARiiON training classes in Irvine CA next month.  I figure it’s about time.  I’ve never had any CLARiiON classes. And this CX4960 is slightly different than my CX700.

Thank you.

6 Operator

 • 

2.1K Posts

December 29th, 2009 08:00

I did some digging but could not come up with confirmation of your theory about the lack of alert because the drive was unused. I'm sure I remember seeing something like that somewhere, but I can't seem to find it now. Maybe someone else can confirm or deny that theory, but it does make some sense.

If you don't get an error before your course next month, you can grill your instructor and see how good they are :-) Hopefully you get a better one than I had (many years ago on the FC4700-2). The training materials were full of mistakes at the time and the instructor would start us on a lab and then leave for 30-40 minutes while we struggled through the faulty procedures *lol*  I've taken several EMC courses since then and they have been significantly better though, so I wouldn't worry too much.

As for being new here, there is no need to ask forgivenes for any percieved ignorance. Most of the folks around here started out asking lots of questions, and over time you will (hopefully) start answering some of the questions from the "new" new folks. Most of the people who have been around for a long time are still asking questions as well, as there is just too much to know about these systems for any one person to know it all. This community gives us all the opportunity to brainstorm, bounce ideas off each other, and occasionally plead for help :-)

30 Posts

December 29th, 2009 10:00

WOW. That is exactly what I experienced.  I guess there is something to be said about reading the “release notes”.    It’s nice to know that it is not dysfunctional.

Thank you very much.

6 Operator

 • 

2.1K Posts

December 29th, 2009 11:00

When I was looking for this I skipped right over the FLARE release notes. I checked the Navisphere admin guides and did couple of searches on the Knowledge Base.

Thanks for digging this up!

34 Posts

December 30th, 2009 05:00

To your question if EMC should respond to your test events, EMC does. If you have email home or dial home configured right, each test even actually generates an SR which is fed in to the SR database. The only quirk is that no engineer is ever alerted of it because the system, knowing that it was a test event, would cancel the SR immediately.

You can actually get in touch with most accessible EMC field engineer and ask him to check the EMC SR database and confirm if your test event has been logged. Field personnel use this method quite effectively to verify if alerting systems are operational.

Hope this helps.

Cheers,

-joji

6 Operator

 • 

2.1K Posts

December 31st, 2009 06:00

You shouldn't even need to contact a field engineer. I've done this before and just called into the regular support line and the first level support folks there can confirm if the dial home was received. We do this if we have to do any work that touches the dial home system/modem/phone line just to confirm that everything is good before we need it.

6 Operator

 • 

2.1K Posts

December 31st, 2009 10:00

Yes, after one incident of a double faulted RAID 5 group (the first failure didn't dial home and we had some issues with our ControlCenter monitoring of the array at the time) we discovered that it is much easier to validate the alerting regularly than to ask "Why didn't it work" after you really needed it.

34 Posts

December 31st, 2009 10:00

Perfecto! That should be the easiest way. It is always a good practice to check alerting systems from time to time.

Cheers,

-joji

No Events found!

Top