lost alternate link

[mobidyc]

Hello,

strange situation, one PV in a VG seems to be inaccessible, strange because it must be an alternate link.

pv in error : c6t1d7
pv ok : c8t1d7
these two dsk are the same disk on a different path, i've other dsk with their alternate links OK, i don't understand the situation.

#> vgdisplay -v vgdata 1>/dev/null

vgdisplay: Warning: couldn't query physical volume "/dev/dsk/c6t1d7":
The specified path does not correspond to physical volume attached to
this volume group
vgdisplay: Warning: couldn't query all of the physical volumes.

#> vgdisplay -v vgdata 2>/dev/null
--- Volume groups ---
VG Name /dev/vgdata
VG Write Access read/write
VG Status available
Max LV 255
Cur LV 32
Open LV 32
Max PV 255
Cur PV 13
Act PV 13
Max PE per PV 2905
VGDA 26
PE Size (Mbytes) 256
Total PE 2765
Alloc PE 1856
Free PE 909
Total PVG 0
Total Spare PVs 0
Total Spare PVs in use 0



PV Name /dev/dsk/c8t1d2
PV Name /dev/dsk/c6t1d2 Alternate Link
PV Status available
Total PE 119
Free PE 61
Autoswitch On

PV Name /dev/dsk/c8t1d7
PV Status available
Total PE 137
Free PE 97
Autoswitch On

#> pvdisplay /dev/dsk/c6t1d7
pvdisplay: Warning: couldn't query physical volume "/dev/dsk/c6t1d7":
The specified path does not correspond to physical volume attached to
this volume group
pvdisplay: Warning: couldn't query all of the physical volumes.
pvdisplay: Couldn't retrieve the names of the physical volumes
belonging to volume group "/dev/vgdata".
pvdisplay: Cannot display physical volume "/dev/dsk/c6t1d7".

#> uname -r
B.11.00

could you help me to understand the problem ?

what i've do:
#> mv /etc/lvmtab /etc/lvmtab.old
#> vgscan -v
Creating "/etc/lvmtab".
vgscan: Couldn't access the list of physical volumes for volume group "/dev/vg00".
vgscan: Couldn't access the list of physical volumes for volume group "/dev/vgdata".
Couldn't stat physical volume "/dev/dsk/c3t2d0":
Invalid argument

/dev/vg00
/dev/dsk/c1t2d0
/dev/dsk/c2t2d0

/dev/vgdata
/dev/dsk/c6t0d0
/dev/dsk/c6t0d2
/dev/dsk/c6t0d3
/dev/dsk/c6t0d4
/dev/dsk/c6t0d5
/dev/dsk/c6t0d6
/dev/dsk/c6t1d0
/dev/dsk/c6t1d1
/dev/dsk/c6t1d2
/dev/dsk/c6t1d3
/dev/dsk/c6t1d5
/dev/dsk/c6t1d6
/dev/dsk/c6t1d7
/dev/dsk/c8t0d0
/dev/dsk/c8t0d2
/dev/dsk/c8t0d3
/dev/dsk/c8t0d4
/dev/dsk/c8t0d5
/dev/dsk/c8t0d6
/dev/dsk/c8t1d0
/dev/dsk/c8t1d1
/dev/dsk/c8t1d2
/dev/dsk/c8t1d3
/dev/dsk/c8t1d5
/dev/dsk/c8t1d6
/dev/dsk/c8t1d7

#> diskinfo /dev/rdsk/c8t1d7
SCSI describe of /dev/rdsk/c8t1d7:
vendor: DGC
product id: CX600WDR5
type: direct access
size: 36048128 Kbytes
bytes per sector: 512

#> diskinfo /dev/rdsk/c6t1d7
SCSI describe of /dev/rdsk/c6t1d7:
vendor: DGC
product id: CX600WDR5
type: direct access
size: 36048128 Kbytes
bytes per sector: 512

#> lssf /dev/rdsk/c8t1d7 /dev/rdsk/c6t1d7
sdisk card instance 8 SCSI target 1 SCSI LUN 7 section 0 at address 0/12/0/0.1.40.0.0.1.7 /dev/rdsk/c8t1d7
sdisk card instance 6 SCSI target 1 SCSI LUN 7 section 0 at address 0/12/0/0.1.13.0.0.1.7 /dev/rdsk/c6t1d7

# ioscan on these two hwpath
Class I H/W Path Driver S/W State H/W Type Description
==================================================================================
disk 47 0/12/0/0.1.40.0.0.1.7 sdisk CLAIMED DEVICE DGC CX600WDR5
/dev/dsk/c8t1d7 /dev/rdsk/c8t1d7
disk 17 0/12/0/0.1.13.0.0.1.7 sdisk CLAIMED DEVICE DGC CX600WDR5
/dev/dsk/c6t1d7 /dev/rdsk/c6t1d7

thanks in advance and sorry for this bad english (difficult to find the good words today) ;)

please help

[Coolmar · Accepted solution]

Hi,

Have a look at this thread:
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1025994

Perhaps the path changed for some reason...did you try an ioscan -funC disk and then insf -e?

[George Liu_4]

verify "/dev/dsk/c6t1d7" is still good

maybe by

dd if="/dev/dsk/c6t1d7" of=/dev/null

[Michael Steele_2]

Dear mobidyc:

Odd things like this happens with disk arrays in a SAN environment so you SAN engineer should be consulted. More then likely the device name changed do to some SAN reprovisioning. And you'll have to do several things to correct this.

a) ioscan (* this will refresh the kernel *)

b) verify new ioscan (* you're looking for a new device with icompletly listed device files. It will only have one line instead of two or more.

c) insf -e (* note any errors and device name changes. If you found a new device then this will automatically generate new device files.

d) at this point, assuming you have a new device, save /etc/lvmtab and regenerate a new lvmtab with vgscan -a -v -c (* -c for every controller, -a for every device *) Chances are you'll still get errors until you complete the next step.

e) vgreduce -f /dev/vg** (* force out the old device

f) vgextend (* add in the new device

g) repeat d)

[John Guster]

If you use EMC NAviAgent, check the agent.config file to see if the tresspass is turned on. If not, update the file and bounce the agent daemon once. Cheers.

[John Guster]

Also, your primary link(path) should be the same path SAN guy assigned to it. Otherwise the paths will keep switching back force assuming you don't use powerpath of EMC.

[mobidyc]

thanks all for your answers.

George Liu:
/dev/dsk/c6t1d7 is also /dev/dsk/c8t1d7, no need to dd this dsk because c8t1d7 is OK

Michael Steele:
i already ask my colleague SAN engineer, all is OK for him.

a,b and c was already made with the same errors ;(

for d, vgscan doen't know the -c flag.
no problem for recreating lvmtab:
#> vgscan -a -v
Creating "/etc/lvmtab".
Couldn't stat physical volume "/dev/dsk/c3t2d0":
Invalid argument

/dev/vg00
/dev/dsk/c1t2d0
/dev/dsk/c2t2d0

/dev/vgdata
/dev/dsk/c6t0d0
/dev/dsk/c6t0d2
/dev/dsk/c6t0d3
/dev/dsk/c6t0d4
/dev/dsk/c6t0d5
/dev/dsk/c6t0d6
/dev/dsk/c6t1d0
/dev/dsk/c6t1d1
/dev/dsk/c6t1d2
/dev/dsk/c6t1d3
/dev/dsk/c6t1d5
/dev/dsk/c6t1d6
/dev/dsk/c6t1d7
/dev/dsk/c8t0d0
/dev/dsk/c8t0d2
/dev/dsk/c8t0d3
/dev/dsk/c8t0d4
/dev/dsk/c8t0d5
/dev/dsk/c8t0d6
/dev/dsk/c8t1d0
/dev/dsk/c8t1d1
/dev/dsk/c8t1d2
/dev/dsk/c8t1d3
/dev/dsk/c8t1d5
/dev/dsk/c8t1d6
/dev/dsk/c8t1d7


Scan of Physical Volumes Complete.
*** LVMTAB has been created successfully.
*** If PV links are configured in the system.
*** Do the following to resync information on disk.
*** #1. vgchange -a y
*** #2. lvlnboot -R

i try your e point now:
#> vgreduce -f /dev/vgdata
skip alternate link /dev/dsk/c6t0d5
skip alternate link /dev/dsk/c6t0d6
skip alternate link /dev/dsk/c6t1d0
skip alternate link /dev/dsk/c6t1d1
skip alternate link /dev/dsk/c6t1d2
vgreduce: Couldn't query physical volume "/dev/dsk/c6t1d7":
The specified path does not correspond to physical volume attached to
this volume group
skip alternate link /dev/dsk/c8t0d0
skip alternate link /dev/dsk/c8t0d2
skip alternate link /dev/dsk/c8t0d3
skip alternate link /dev/dsk/c8t0d4
skip alternate link /dev/dsk/c8t1d3
skip alternate link /dev/dsk/c8t1d5
skip alternate link /dev/dsk/c8t1d6

at this point, c6t1d7 is still in the VG

repeat the d point give me the same output...with c6t1d7 again

John Guster:
AutoTrespass is activated
#> grep AutoTrespass /opt/Navisphere/bin/agent.config
OptionsSupported AutoTrespass

Coolmar:
i'm going to read your link

[mobidyc]

problem solved.

#> vgreduce -l vgdata /dev/dsk/c6t1d7
Volume group "vgdata" has been successfully reduced.
Volume Group configuration for /dev/vgdata has been saved in /etc/lvmconf/vgdata.conf

#> vgextend vgdata /dev/dsk/c6t1d7
Volume group "vgdata" has been successfully extended.
Volume Group configuration for /dev/vgdata has been saved in /etc/lvmconf/vgdata.conf

#> vgdisplay -v vgdata


PV Name /dev/dsk/c8t1d7
PV Name /dev/dsk/c6t1d7 Alternate Link
PV Status available
Total PE 137
Free PE 97
Autoswitch On

all seems OK now

thanks all for your help.

[Michael Steele_2]

Dear mobidyc:

It seems most odd to me that c6t1d7 didn't have a device file renamed. What exactly was the source of the problem? A new lun id perhaps? Its got to be associated to the SAN.