A workaround for hacmp when Aix suddenly loses power

Aix is a Unix-like operating system, and HACMP is a control application for AIX systems, when Aix suddenly power off, there will be hacmp can not start the situation, the following small series to introduce the Aix after power off HACMP can not start the solution.

Because UPS problem causes the database minicomputer and the storage and so on power off, the hacmp cannot start,

Found JZPT1 on the hard disk identification problem, the original hard drive is from HDISK0--HDISK11, now is hdisk2 to Hdisk9 No, behind from the hdisk12 after a lot of hard drives, through the RMDEV-RDL

Hdisk2 to all the end of the hard drive, Cfgmgr-v again, but in the IMPORTVG times wrong,

0516-082 lqueryvg:unable to access a special device file.

Execute REDEFINEVG and SYNCLVODM to build correct environment.

0516-082 lqueryvg:unable to access a special device file.

Execute REDEFINEVG and SYNCLVODM to build correct environment.

0516-1140 importvg:unable to read the volume Group descriptor area

On specified physical volume.

None of the newly identified hard drives are Pvid

　　Use the following command:

Chdev-l hdisk2-a Pv=yes, Error:

Method Error (/usr/lib/methods/chgdisk):

0514-047 cannot access a device.

　　Solve the problem using the following methods:

RMDEV-RDL Hdisk2

RMDEV-RDL HDISK3

。。。。

RMDEV-RDL hdisk11

Remove all disks

RMDEV-RDL Fcs0

。。。。

RMDEV-RDL FCS3

Cfgmgr-v

All PV was found to have Pvid

But found

The HDISK2-HDISK8 on the JZPT1 is consistent with the PV sequence on the Jzpt2, but HDISK9 (JZPT2), which also belongs to Datavg, is jzpt1. There are two hard drives in the same order,

And it's all datavg. It seems that in the future when doing shared VG, try not to use multiple small LUNs, but with a small number of large capacity LUNs, in order to reduce this hdisk order inconsistent imagination.

Had to, in Jzpt2 Admiral Datavg Varyon, then Smitty REDUCEVG, but the hint said will destroy the Datavg on the bare device content.

Had to use the Smitty MIGRATEPV command to move the contents of the HDISK9 to Hdisk2. After turning lspv-p Datavg, found that there is no longer content on the HDISK9, and hdisk2 content than HDISK3

More disk content.

Lsvg-p Datavg

DATAVG:

Pv_name PV State Total pps free PPS free distribution

Hdisk2 Active 1599 811 62..00..109..320..320

Hdisk3 Active 1599 1155 192..04..319..320..320

Hdisk4 Active 1599 1155 192..04..319..320..320

Hdisk5 Active 1599 1087 64..64..319..320..320

Hdisk6 Active 1599 1087 64..64..319..320..320

Hdisk7 Active 1599 1135 112..64..319..320..320

Hdisk8 Active 1599 1215 192..64..319..320..320

HDISK9 Active 1599 1599 192..64..319..320..320

Again using Smitty REDUCEVG will HDISK9 from Datavg removed. So the hdisk in the datavg of JZPT1 and Jzpt2 is consistent.

Perform Smitty IMPORTVG on JZPT1

VG writes DATAVG,PV write to Hdisk2. Datavg was imported. Assign the bare devices in the Datavg to the ORACLE:DBA user group.

Problem solving

Later found that jzpt1 hacmp can start, datavg can also pull up, but Jzpt2 datavg, so the jzpt2 on all the storage disk deleted, remove the HBA card, and then again

The underlying problem is that the operating system does not have the storage multipath software installed.

Above is the method that the Aix hacmp cannot start after the power outage, in order to avoid similar situation again, it is best to install the storage multipath software in the AIX system.