Or at least, one I’ve not seen before, anyway, error below, does anyone know what it’s referring to?
fmmb.BlobNotFound: blob_id="3", owner="PARTNER", reason="no opaque data type"
Thanks,
~Mark
mark.flint@sanger.ac.uk
According to syslog translator
EMS Identifier
fmmb.BlobNotFound
Syslog Message
Not Available
Severity
WARNING
Description
This message occurs when a mailbox 'blob' is not read from the on-disk mailbox data structure. A 'blob' is a piece of data which has been stored for a non-High Availability (HA) Data ONTAP(R) subsystem in the mailbox.
Corrective Action
If the blob id is "3" and this is a 2-node cluster, use the 'cluster ha modify -configure' command to unconfigure ha and then reconfigure it.
From: toasters-bounces@teaparty.net [mailto:toasters-bounces@teaparty.net] On Behalf Of Mark Flint Sent: Wednesday, May 06, 2015 5:19 PM To: Toasters Subject: New error message
Or at least, one I’ve not seen before, anyway, error below, does anyone know what it’s referring to?
fmmb.BlobNotFound: blob_id="3", owner="PARTNER", reason="no opaque data type"
Thanks,
~Mark
mark.flint@sanger.ac.ukmailto:mark.flint@sanger.ac.uk
Hey Mark,
For cluster-mode/cDOT, and releases before 8.3 this warning can be an indication of something severe in case of 2-node clusters. The mailboxes are used to validate RDB revision in 2-node clusters with HA enabled (in lieu of quorum voting). This message can mean that that communication is out of sync.
This means that takeovers in a 2-node cluster can cause a loss of cluster quorum, so a full loss of data access when taking over in either direction in the ha pair.
Check cluster ha show under set diag.
If it's a 2-node cDOT cluster, and (set diag!) cluster ha show does not state true and healthy for both values (in particular the MBX part, which is only visible in set diag), disable and reenable cluster ha.
i.e. run cluster ha modify -enabled false, and then true (don't have a sim at hand to validate my syntax for you) should fix this.
If it's not a 2-node cluster, a to/gb will likely be enough. Beyond that, open a ticket with support.
This type of scenario can occur if you move the root aggregate via the support site kb article and skip a step. I believe this gets self-corrected in 8.3.
Maarten
On 05/06/2015 07:43 AM, Borzenkov, Andrei wrote:
According to syslog translator
EMS Identifier
fmmb.BlobNotFound
Syslog Message
Not Available
Severity
WARNING
Description
This message occurs when a mailbox 'blob' is not read from the on-disk mailbox data structure. A 'blob' is a piece of data which has been stored for a non-High Availability (HA) Data ONTAP(R) subsystem in the mailbox.
Corrective Action
If the blob id is "3" and this is a 2-node cluster, use the 'cluster ha modify -configure' command to unconfigure ha and then reconfigure it.
*From:*toasters-bounces@teaparty.net [mailto:toasters-bounces@teaparty.net] *On Behalf Of *Mark Flint *Sent:* Wednesday, May 06, 2015 5:19 PM *To:* Toasters *Subject:* New error message
Or at least, one I’ve not seen before, anyway, error below, does anyone know what
it’s referring to?
fmmb.BlobNotFound: blob_id="3", owner="PARTNER", reason="no opaque data type"
Thanks,
~Mark
mark.flint@sanger.ac.uk mailto:mark.flint@sanger.ac.uk
Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters