How to determine the disk name referred to by SCSI or vxdmp messages in /var/adm/messages
Article ID: 100013142
Updated On:
Description
Error Message
Sep 10 15:32:36 superman vxdmp: [ID 997040 kern.notice] NOTICE: vxvm:vxdmp: disabled path 32/0x1cd48 belonging to the dmpnode 201/0x2100Resolution
Extract fromthe /var/adm/messages file:
Sep 16 15:41:59fcis13 scsi: [ID 107833 kern.warning] Warning:/pci@1f,4000/fibre-channel@5/sd@1f,f2 (sd78):
Sep 16 15:41:59fcis13 SCSI transport failed: reason 'tran_err': giving up
Sep 16 15:41:59fcis13 vxdmp: [ID 619769 kern.notice] NOTICE: vxdmp: Path failure on32/0x274
Sep 16 15:41:59fcis13 vxdmp: [ID 997040 kern.notice] NOTICE: vxvm:vxdmp: disabled path 32/0x270belonging to the dmpnode 221/0x28
Option1:
Using SCSIerror messages in /var/adm/messages
Sep 1615:41:59 fcis13 scsi: [ID 107833 kern.warning] Warning:/pci@1f,4000/fibre-channel@5/sd@1f,f2 (sd78):
To identify thedisk path if there are SCSI errors, just convert the SCSI error from hexto dec.
/pci@1f,4000/fibre-channel@5/sd@1f,f2(sd78)
e.g. Useadb:
#echo "0x1f=D" | adb 31 #echo "0xf2=D" | adb 242 Disk/Pathfailure from an SD driver view is found to be: c5t31d242s2
Option2:
Using vxdmpkernel notice in /var/adm/messages
Sep 1615:41:59 fcis13 vxdmp: [ID 619769 kern.notice] NOTICE: vxdmp: Path failure on32/0x274
look at thels -lL /dev/rdsk/* and look for the major number eg. 32. (note that alower case L is the character before the upper case L)
Then convertthe 0x274 to decimal
#echo "0x274=D" | adb 628 So, majornumber 32 and minor number 628 need to be located
#ls -lL /dev/rdsk/* | grep 628 crw-r-----1 root sys 32,628 Aug 21 13:44 c5t31d242s4 This happenedto be the Volume Manager public region where the disk failure wasnoticed.
OPTION3:
Anotherapproach/procedure similar to the above is as follows:
From/var/adm/messages:
Sep 10 15:32:32superman scsi: [ID 107833 kern.warning] Warning:/pci@5d,600000/pci@1/fibre-channel@4/sd@78,3c (sd14761):
Sep 10 15:32:32superman SCSI transport failed: reason 'tran_err': retrying command
Sep 10 15:32:32superman scsi: [ID 107833 kern.warning] Warning:/pci@5d,600000/pci@1/fibre-channel@4/sd@78,3c (sd14761):
Sep 10 15:32:32superman SCSI transport failed: reason 'tran_err': giving up
Sep 10 15:32:32superman scsi: [ID 107833 kern.warning] Warning:/pci@5d,600000/pci@1/fibre-channel@4/sd@78,3c (sd14761):
Sep 10 15:32:32superman SCSI transport failed: reason 'tran_err': giving up
Sep 10 15:32:32superman scsi: [ID 107833 kern.warning] Warning:/pci@5d,600000/pci@1/fibre-channel@4/sd@78,3c (sd14761):
Sep 10 15:32:32superman SCSI transport failed: reason 'tran_err': giving up
Sep 10 15:32:32superman scsi: [ID 107833 kern.warning] Warning:/pci@5d,600000/pci@1/fibre-channel@4/sd@78,3c (sd14761):
Sep 10 15:32:32superman SCSI transport failed: reason 'tran_err': giving up
Sep 10 15:32:32superman scsi: [ID 107833 kern.warning] Warning:/pci@5d,600000/pci@1/fibre-channel@4/sd@78,3c (sd14761):
Sep 10 15:32:32superman SCSI transport failed: reason 'tran_err': giving up
Sep 10 15:32:32superman scsi: [ID 107833 kern.warning] Warning:/pci@5d,600000/pci@1/fibre-channel@4/sd@78,3c (sd14761):
Sep 10 15:32:32superman SCSI transport failed: reason 'tran_err': giving up
Sep 10 15:32:32superman scsi: [ID 107833 kern.warning] Warning:/pci@5d,600000/pci@1/fibre-channel@4/sd@78,3c (sd14761):
Sep 10 15:32:32superman offline
Sep 10 15:32:36superman vxdmp: [ID 619769 kern.notice] NOTICE: vxdmp: Path failure on32/0x1cd4c
Sep 10 15:32:36superman vxdmp: [ID 997040 kern.notice] NOTICE: vxvm:vxdmp: disabled path32/0x1cd48 belonging to the dmpnode 201/0x2100
Sep 10 15:32:38superman scsi: [ID 365881 kern.notice]
You can seethat the first vxdmp messages is the "disabled path 32/0x1cd48 belonging to thedmpnode 201/0x2100"
VxDMP has Xnumber of paths depending on your setup.
If a pathbecomes disabled you will see "disabled path" error. This error message meansthat one of the X number of paths are down for whatever reason.
You don'twant to focus on the path (disabled path 32/0x1cd48), but rather the dmpnode(dmpnode 201/0x2100) in finding the c#t#d#.
We can use the/var/adm/messages to find out the DMPnode which is the entire metadevice(c#t#d#) failed by the following:
"disabled path32/0x1cd48 belonging to the dmpnode 201/0x2100"
1. 201is the major number of the VxDMP driver.
#grep vxdmp/etc/name_to_major
vxdmp201
2.0x2100 is the Hexadecimal number which needs to be converted toDecimal.
0x2100 hex =8448 dec
3. Use the/dev/vx/dmp tree to find the c#t#d#.
#ls -la/dev/vx/dmp |grep 8448
brw------- 1root root 201,8448 Sep 27 10:18 c10t120d60s0
So the paththat failed is c10t120d60.
4. To finds italternative path use vxdmpadm.
/usr/sbin/vxdmpadmgetsubpaths dmpnodename=c10t120d60s2
/usr/sbin/vxdmpadmgetsubpaths dmpnodename=c10t120d60s2
NAME STATEPATH-TYPE CTLR-NAME ENCLR-TYPE ENCLR-NAME
====================================================================
c10t120d60s2DISABLED - c10 EMC emc3
c17t121d60s2ENABLED - c17 EMC emc3
To conclude,the dmpnode 201/0x2100 found in /var/adm/messages is related to metanodec10t120d60 which has two paths to the drive (c10t120d60 andc17t121d60)
Issue/Introduction
How to determine the disk name referred to by SCSI or vxdmp messages in /var/adm/messages
Was this article helpful?
Yes
No
Powered by
