topic Re: mirror disk error in Operating System - HP-UX

mirror disk error

emily_3 — Fri, 02 May 2003 08:07:51 GMT

hello, My server have two harddisk, and everything is in the 1th. disk.
I make the mirror LVM for lvol1 (/stand) lvol2 (/swap) lvol3 (/) lvol4 (/tmp) from 1th. disk to the 2th. disk.
then I showdown and pull out the 1th. disk, and tried to boot from the 2th. disk.
However, it boot up with error, and i found that /etc/services file is overwriten, and the size is 0.
If I am not wrong, /etc diretory should be under the /(root) LVM, correct? and I have mirror this LVM to the 2th. disk.
Can anybody help to explain??

Re: mirror disk error

Michael Tully — Fri, 02 May 2003 08:17:19 GMT

What was the date/time of the /etc/services file? Before or after boot time...
What error(s) did you encounter when attempting to boot from your mirrored disk?

Re: mirror disk error

emily_3 — Fri, 02 May 2003 08:39:15 GMT

hi, should be before boot up, because the error during boot up is due to /etc/services.
here is the /var/adm/syslog/syslog.log error message:

May 2 01:21:12 ssrvr212 rpcbind: check_netconfig: Found CLTS loopback transport
May 2 01:21:12 ssrvr212 rpcbind: check_netconfig: Found COTS loopback transport
May 2 01:21:12 ssrvr212 rpcbind: check_netconfig: Found COTS ORD loopback transport
May 2 01:21:12 ssrvr212 rpcbind: init_transport: check binding for udp
May 2 01:21:12 ssrvr212 rpcbind: init_transport: check binding for tcp
May 2 01:21:12 ssrvr212 rpcbind: init_transport: check binding for ticlts
May 2 01:21:12 ssrvr212 rpcbind: init_transport: check binding for ticotsord
May 2 01:21:12 ssrvr212 rpcbind: init_transport: check binding for ticots
May 2 01:21:12 ssrvr212 rpcbind: init_transport: ticlts: address in use
May 2 01:21:12 ssrvr212 rpcbind: init_transport: ticots: address in use
May 2 01:21:12 ssrvr212 rpcbind: init_transport: ticotsord: address in use
May 2 01:21:15 ssrvr212 inetd[637]: Reading configuration
May 2 01:21:15 ssrvr212 inetd[637]: ftp/tcp: Unknown service
May 2 01:21:15 ssrvr212 inetd[637]: telnet/tcp: Unknown service
May 2 01:21:15 ssrvr212 inetd[637]: login/tcp: Unknown service
May 2 01:21:15 ssrvr212 inetd[637]: shell/tcp: Unknown service
May 2 01:21:15 ssrvr212 inetd[637]: exec/tcp: Unknown service
May 2 01:21:15 ssrvr212 inetd[637]: ntalk/udp: Unknown service
May 2 01:21:15 ssrvr212 inetd[637]: ident/tcp: Unknown service
May 2 01:21:15 ssrvr212 inetd[637]: printer/tcp: Unknown service
May 2 01:21:15 ssrvr212 inetd[637]: daytime/tcp: Unknown service
May 2 01:21:15 ssrvr212 inetd[637]: daytime/udp: Unknown service
May 2 01:21:15 ssrvr212 inetd[637]: time/tcp: Unknown service
syslog.log (65%)

Re: mirror disk error

T G Manikandan — Fri, 02 May 2003 08:57:32 GMT

Are you using nis or nis+.

If that is the case then how does your /etc/nsswitch.conf file look like.
Does it point to that and not to /etc/services file.
ALso check whether the /etc/nsswitch.conf file is present.

revert

Re: mirror disk error

emily_3 — Fri, 02 May 2003 09:19:34 GMT

I checked that /etc/nsswith.conf is not exist.
Thanks for your help.

I need to leave now, if you need any more information, please post it to me. I will reply you next week.Thanks.

Re: mirror disk error

RolandH — Thu, 15 May 2003 11:44:37 GMT

Hi Emily,

you have took this question in the ignite-ux group, so I must ask you - Have you installed this machine with a golden image or a recovery tape? Ignite normally not recover mirros. You must redo your mirror.
So boot from your pri disk and check, if your lvol has a mirror copy.
# lvdisplay /dev/vg00/lvol3
--- Logical volumes ---
LV Name /dev/vg00/lvol3
VG Name /dev/vg00
LV Permission read/write
LV Status available/syncd
Mirror copies 1
Consistency Recovery MWC
Schedule parallel
LV Size (Mbytes) 26208
Current LE 819
Allocated PE 1638
Stripes 0
Stripe Size (Kbytes) 0
Bad block off
Allocation strict/contiguous
IO Timeout (Seconds) default

Your "Mirror copies" value must be >0.
If you have mirror copies then check, if you have PE that are in status syncd. So your "LV Status" is = available/syncd.
If not check the state of the PE with:
# lvdisplay -v /dev/vg00/lvol3 ( you see somthing like that )
--- Logical volumes ---
LV Name /dev/vg00/lvol3
VG Name /dev/vg00
LV Permission read/write
LV Status available/syncd
Mirror copies 1
Consistency Recovery MWC
Schedule parallel
LV Size (Mbytes) 26208
Current LE 819
Allocated PE 1638
Stripes 0
Stripe Size (Kbytes) 0
Bad block off
Allocation strict/contiguous
IO Timeout (Seconds) default

--- Distribution of logical volume ---
PV Name LE on PV PE on PV
/dev/dsk/c1t2d0 819 819
/dev/dsk/c2t2d0 819 819

--- Logical extents ---
LE PV1 PE1 Status 1 PV2 PE2 Status 2
00000 /dev/dsk/c1t2d0 00266 current /dev/dsk/c2t2d0 00266 current
00001 /dev/dsk/c1t2d0 00267 current /dev/dsk/c2t2d0 00267 current
00002 /dev/dsk/c1t2d0 00268 current /dev/dsk/c2t2d0 00268 current
...

The PE are in state "current". That ok. If they some or all are "stale" do a sync.
# sync

Kindly regards
Roland