dynamox

updated

13 years ago

D

dynamox

11 Legend

•

20419 Posts

•

87439 Points

0

5056

October 4th, 2013 07:00

Boot mirror is critical

Looks like i have an issue with boot drive(s) in my 108NL node.

10-4-2013 10-53-02 AM.png

What do i need to do ? Will the box call-home and open an SR (supportIQ is enabled) or do i have to call EMC ? What it's involved in replacing the failed drive ?

Thank you

  • Peter_Sero

    6 Operator

    •

    1169 Posts

    2478

    0

    Posted October 4th, 2013 08:00

    Such things I usually inquire via

    https://support.emc.com/servicecenter/liveChat/

    The docs alone

    https://support.emc.com/docu44208_Replace-a-Boot-Drive-in-a-108NL-Node.pdf

    wouldn't help without a new drive in your hands...

    -- Peter

  • dynamox

    11 Legend

    •

    20419 Posts

    •

    87439 Points

    2478

    0

    Posted October 4th, 2013 13:00

    Thanks Peter, this is what i got. EMC folks ..should this automatically generate a ticket via email-home ? Something that should be monitored and alerted to EMC automatically.

    # gmirror status

                     Name    Status  Components

             mirror/root1  DEGRADED  ad2p4

         mirror/var-crash  COMPLETE  ad3p9

               mirror/mfg  COMPLETE  ad3p8

                                     ad2p9

    mirror/journal-backup  COMPLETE  ad3p7

                                     ad2p7

              mirror/var1  DEGRADED  ad2p6

              mirror/var0  DEGRADED  ad2p5

             mirror/root0  DEGRADED  ad2p3

  • dynamox

    11 Legend

    •

    20419 Posts

    •

    87439 Points

    2478

    0

    Posted October 4th, 2013 14:00

    SR was created last night,  i should have checked prior.

    Thank you

  • Peter_Sero

    6 Operator

    •

    1169 Posts

    2478

    0

    Posted October 4th, 2013 14:00

  • Peter_Sero

    6 Operator

    •

    1169 Posts

    2478

    2

    Posted October 4th, 2013 14:00

    Alerts have been created, which is good. First, check the numerical alert type:

    isilon-1# isi alert show 8.34554                                             

              ID: 8.34554

            Type: 400100004

    ...

    ...

    ...

    This example has type 400100004 (different -- not your mirror issue.)

    Next, review the notification rules:

    isilon-1# isi alert notifications ls                                         

    NAME                 TYPE      RECIPIENTS DESCRIPTION                 

    Isilon support email smtp      1          9 categories, 19 event types

    SupportIQ            supportiq 1          9 categories, 1 event types 

    site-itsupport          smtp      1          9 categories, 152 event types

    Here, two rules reach out to EMC/isilon, "Isilon support email" and "SupportIQ".

    Finally check wether the notification rules cover the specific alert type:

    isilon-1# isi alert notifications ls -n'Isilon support email' | grep 400100004

       C    SW_JOBENG_JOB_FAILED (400100004)

    isilon-1# isi alert notifications ls -n'SupportIQ' | grep 400100004

    (empty output)

    This means, in the example, alert type 400100004 is configured to be sent out as C=Critical via

    e-mail/smtp due to the first rule; but not due to the built-in SupportIQ rule!

    Note: You can check the configured alerts for e-mail notif. rules in the WebGUI,

    but not for the SupportIQ rule. The latter can be inspected only on the CLI.

    Of course, the cluster e-mail config should actually work for

    sending out mails to EMC...

    It's a long chain after all:

    HW/SW issue -> raise alert -> trigger notif. -> send smtp mail -> EMC receives -> generate SR (or not, depends on alert type/severity) -> cluster admin gets notified

    I would double-check directly with EMC support for the SR.

    -- Peter