Hi Steve.
I've just checked our (non-Metro) clusters and I can confirm that most
(probably all) the volumes that contain LUNs have nvfail set to 'on.'
Certainly there are no 'non-LUN' volumes with nvfail set to on.
Our transition 'through' 9.8 was directly from 9.7P13 to 9.9.1P7.
We haven't had any LUNs go offline but we also haven't had any FO/GB events
since the last ONTAP upgrade.
HTH. Please update the list if you find out more.
Kind regards
Steve
On Fri, 24 Jun 2022 at 14:52, Steve Kirk via Toasters <toasters@teaparty.net>
wrote:
>
>
>
> ---------- Forwarded message ----------
> From: Steve Kirk <steve.kirk@sheffield.ac.uk>
> To: toasters@teaparty.net
> Cc:
> Bcc:
> Date: Fri, 24 Jun 2022 14:48:44 +0100
> Subject: NVFAIL - new behaviour or setting?
> Hello toasters,
>
> I'm interested to know if anyone else has found recently that volumes
> containing LUNs have gone into the 'NVFAIL' state during an ONTAP
> upgrade or other event that causes an HA failover?
>
> It's not something we have ever specified on volume creation;
> documentation suggests that it should be disabled by default however
> all of our volumes containing LUNs appear to have NVFAIL enabled. My
> *perception* this is a change since we upgraded to ONTAP 9.8.
>
> The documentation seems live under Metrocluster; our affected system
> is not a Metrocluster:
>
> https://docs.netapp.com/us-en/ontap-metrocluster/manage/concept_monitoring_and_protecting_database_validity_by_using_nvfail.html#how-nvfail-impacts-access-to-nfs-volumes-or-luns
>
> It's also odd as the KB article says:
> - This affords no extra protection for Storage Area Network (SAN)
> hosts because SAN hosts reference LUNs by the LUN number assigned via
> LUN maps and NOT by the LUN name.
> - This protection was designed for file protocols, such as Network
> File System (NFS) and Common Internet File System protocol (CIFS),
> protect against corruption due to NVFAIL.
>
> Taken from here:
>
> https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Software/ONTAP_OS/How_does_NVFAIL_work_with_LUNs_in_clustered_Data_ONTAP
>
> So it also seems like either the setting or the documentation is the
> wrong way round to me...
>
> Looking for feedback to see if anyone else has experienced this or if
> there's something obvious we've missed before I raise a support call.
>
> Cheers,
> Steve
>
>
>
> ---------- Forwarded message ----------
> From: Steve Kirk via Toasters <toasters@teaparty.net>
> To: toasters@teaparty.net
> Cc:
> Bcc:
> Date: Fri, 24 Jun 2022 14:49:13 +0100
> Subject:
> _______________________________________________
> Toasters mailing list
> Toasters@teaparty.net
> https://www.teaparty.net/mailman/listinfo/toasters
I've just checked our (non-Metro) clusters and I can confirm that most
(probably all) the volumes that contain LUNs have nvfail set to 'on.'
Certainly there are no 'non-LUN' volumes with nvfail set to on.
Our transition 'through' 9.8 was directly from 9.7P13 to 9.9.1P7.
We haven't had any LUNs go offline but we also haven't had any FO/GB events
since the last ONTAP upgrade.
HTH. Please update the list if you find out more.
Kind regards
Steve
On Fri, 24 Jun 2022 at 14:52, Steve Kirk via Toasters <toasters@teaparty.net>
wrote:
>
>
>
> ---------- Forwarded message ----------
> From: Steve Kirk <steve.kirk@sheffield.ac.uk>
> To: toasters@teaparty.net
> Cc:
> Bcc:
> Date: Fri, 24 Jun 2022 14:48:44 +0100
> Subject: NVFAIL - new behaviour or setting?
> Hello toasters,
>
> I'm interested to know if anyone else has found recently that volumes
> containing LUNs have gone into the 'NVFAIL' state during an ONTAP
> upgrade or other event that causes an HA failover?
>
> It's not something we have ever specified on volume creation;
> documentation suggests that it should be disabled by default however
> all of our volumes containing LUNs appear to have NVFAIL enabled. My
> *perception* this is a change since we upgraded to ONTAP 9.8.
>
> The documentation seems live under Metrocluster; our affected system
> is not a Metrocluster:
>
> https://docs.netapp.com/us-en/ontap-metrocluster/manage/concept_monitoring_and_protecting_database_validity_by_using_nvfail.html#how-nvfail-impacts-access-to-nfs-volumes-or-luns
>
> It's also odd as the KB article says:
> - This affords no extra protection for Storage Area Network (SAN)
> hosts because SAN hosts reference LUNs by the LUN number assigned via
> LUN maps and NOT by the LUN name.
> - This protection was designed for file protocols, such as Network
> File System (NFS) and Common Internet File System protocol (CIFS),
> protect against corruption due to NVFAIL.
>
> Taken from here:
>
> https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Software/ONTAP_OS/How_does_NVFAIL_work_with_LUNs_in_clustered_Data_ONTAP
>
> So it also seems like either the setting or the documentation is the
> wrong way round to me...
>
> Looking for feedback to see if anyone else has experienced this or if
> there's something obvious we've missed before I raise a support call.
>
> Cheers,
> Steve
>
>
>
> ---------- Forwarded message ----------
> From: Steve Kirk via Toasters <toasters@teaparty.net>
> To: toasters@teaparty.net
> Cc:
> Bcc:
> Date: Fri, 24 Jun 2022 14:49:13 +0100
> Subject:
> _______________________________________________
> Toasters mailing list
> Toasters@teaparty.net
> https://www.teaparty.net/mailman/listinfo/toasters