AnsweredAssumed Answered

M3 CLDB failure - "MissingDisk"

Question asked by thealy on Jun 24, 2013
Latest reply on Jun 25, 2013 by thealy
M3 cluster running Version: 2.1.3.19871.GA

Log search eventually found CLDB failure due to "Missing Disk". The CLDB node has/had 5 identical 750GB drives, but disktab looks like this: (/dev/sda is the system disk)

/dev/sdb 5A1D3D8C-D863-2C40-B459-0C067AB35000
/MissingDisk1 434E2D3A-AE44-ED71-8A2A-0F067AB35000
/dev/sdc 0AEEDD11-D613-87F5-983C-03C002635100
/dev/sdd D2BC19C8-10B2-8740-8B51-04C002635100

(Where presumably the missing disk is /dev/sde)

A forum search suggested running this:

/opt/mapr/server/mrconfig disk list
ListDisks resp: status 0 count=3

ListDisks /dev/sdb
DG 0: Single SingleDisk5 Online
DG 1: Raid0 Stripe42 Offline
DG 2: Concat Concat43 Offline
SP 0: name SP4, Offline, size 1429758 MB, free 0 MB, path /dev/sdb
ListDisks /dev/sdc
DG 0: Single SingleDisk5 Online
DG 1: Raid0 Stripe32 Online
DG 2: Concat Concat33 Online
SP 0: name SP3, Online, size 1370774 MB, free 1164004 MB, path /dev/sdc
ListDisks /dev/sdd
DG 0: Single SingleDisk5 Online
DG 1: Raid0 Stripe32 Online
SP 0: name SP3, Online, size 1370774 MB, free 1164004 MB, path /dev/sdc

And also this recommendation which I tried to no avail:

pls comment this line

/MissingDisk9 [disk id]

in conf/disktab

and run

/opt/mapr/server/mrconfig sp refresh

On any other node I would put in a spare disk and fire it up. But I'm not sure if this is advisable on the CLDB. Full power-cycle of the system did not return the disk.

Suggestions?

Outcomes