AIX ODM for MPIO User Guide 09 -

This chapter provides answers to frequently asked questions (FAQs) during AIX ODM for MPIO configuration and management. You can refer to this chapter for troubleshooting if necessary.

7.1 What Are Known Bugs That Affect This Software and Corresponding Patches?

Question

What are known bugs that affect this software and corresponding patches?

Answer

The following describes the known system bugs that may cause SAN Boot failure and corresponding patches:

Bug description:

A REBOOT MAY RESULT IN RANDOM MPIO PATHS IN A MISSING STATE

Patch information:
5300-08 maps to patch IZ66562; 5300-09 maps to patch IZ64005; 5300-10 maps to patch IZ64011; 5300-11 maps to patch IZ65325; 5300-12 maps to patch IZ65775; 6100-01 maps to patch IZ71807; 6100-02 maps to patch IZ64000; 6100-03 maps to patch IZ63159; 6100-04 maps to patch IZ64133; 6100-05 maps to patch IZ65870; 6100-06 maps to patch IZ66023.

7.2 Modifying the Attributes of the fscsi Device

Question

How do you modify the dyntrk and fc_err_recov parameters for the fscsi device?

Answer

There are three methods to modify the dyntrk and fc_err_recov parameters of the fscsi device. Select a proper method based on actual conditions.

Method 1
If the host can be restarted, use method 1 to modify the parameters.
1. Run the chdev -l fscsiX -a dyntrk=yes -P and chdev -l fscsiX -a fc_err_recov=fast_fail -P commands to modify the attributes of all HBAs that connect to switches.
```
-bash-3.00# chdev -l fscsi0 -a dyntrk=yes -P
fscsi0 changed
-bash-3.00# chdev -l fscsi0 -a fc_err_recov=fast_fail -P
fscsi0 changed
```
2. Restart the host to make the modification take effect.
Method 2
If the fscsi device connected to Huawei storage has no virtual disks or has virtual disks that can be deleted, use method 2. If it has virtual disks that cannot be deleted or may not be deleted, use method 3 instead.
1. Run the rmdev -l fscsiX -R command to set the states of all fscsi devices connected to Huawei storage to defined.
```
-bash-3.00# rmdev -l fscsi0 -R
fscsi0 Defined
```
2. Run the chdev -l fscsiX -a dyntrk=yes and chdev -l fscsiX -a fc_err_recov=fast_fail commands to modify the attributes of all fscsi devices connected to Huawei storage.
```
-bash-3.00# chdev -l fscsi0 -a dyntrk=yes
fscsi0 changed
-bash-3.00# chdev -l fscsi0 -a fc_err_recov=fast_fail
fscsi0 changed
```
3. Run the cfgmgr -l fscsiX command to reconfigure all fscsi devices connected to Huawei storage.
```
-bash-3.00# cfgmgr -l fscsi0
-bash-3.00#
```
Method 3
If the reported virtual disks of Huawei storage are connected to multiple fscsi devices, use method 3 (no virtual disks need to be deleted). If there is only one fscsi device connected to Huawei storage, use method 1 or 2.
1. Run the lspath command to confirm the fscsi device in use. In the following figure, hdisk1 and hdisk2 are virtual disks of Huawei storage, which are connected to fscsi0 and fscsi1.
```
-bash-3.00# lspath
Enabled hdisk0 scsi0
Enabled hdisk1 fscsi0
Enabled hdisk2 fscsi0
Enabled hdisk1 fscsi0
Enabled hdisk2 fscsi0
Enabled hdisk1 fscsi1
Enabled hdisk2 fscsi1
Enabled hdisk1 fscsi1
Enabled hdisk2 fscsi1
```
  NOTE:
  The previous screen message indicates that hdisk1 and hdisk2 are virtual disks of Huawei storage, and they are connected to fscsi0 and fscsi1.
2. Run the rmdev -l fscsi0 -R command to set the state of fscsi0 to defined.
```
-bash-3.00# rmdev -l fscsi0 -R
fscsi0 Defined
```
3. Run the chdev -l fscsi0 -a dyntrk=yes and chdev -l fscsi0 -a fc_err_recov=fast_fail commands to modify the attributes of fscsi0.
```
-bash-3.00# chdev -l fscsi0 -a dyntrk=yes
fscsi0 changed
-bash-3.00# chdev -l fscsi0 -a fc_err_recov=fast_fail
fscsi0 changed
```
4. Run the cfgmgr -l fscsi0 command to reconfigure fscsi0.
```
-bash-3.00# cfgmgr -l fscsi0
-bash-3.00#
```
5. Run the lspath command to ensure that all disk paths are normal.
```
-bash-3.00# lspath
Enabled hdisk0 scsi0
Enabled hdisk1 fscsi0
Enabled hdisk2 fscsi0
Enabled hdisk1 fscsi0
Enabled hdisk2 fscsi0
Enabled hdisk1 fscsi1
Enabled hdisk2 fscsi1
Enabled hdisk1 fscsi1
Enabled hdisk2 fscsi1
```
6. Run the rmdev -l fscsi1 -R command to set the state of fscsi1 to defined.
```
-bash-3.00# rmdev -l fscsi1 -R
fscsi1 Defined
```
7. Run the chdev -l fscsi1 -a dyntrk=yes and chdev -l fscsi1 -a fc_err_recov=fast_fail commands to modify the attributes of fscsi1.
```
-bash-3.00# chdev -l fscsi1 -a dyntrk=yes
fscsi1 changed
-bash-3.00# chdev -l fscsi1 -a fc_err_recov=fast_fail
fscsi1 changed
```
8. Run the cfgmgr -l fscsi1 command to reconfigure fscsi1.
```
-bash-3.00# cfgmgr -l fscsi1
-bash-3.00#
```
9. Run the lspath command to ensure that all disk paths are normal.
```
-bash-3.00# lspath
Enabled hdisk0 scsi0
Enabled hdisk1 fscsi0
Enabled hdisk2 fscsi0
Enabled hdisk1 fscsi0
Enabled hdisk2 fscsi0
Enabled hdisk1 fscsi1
Enabled hdisk2 fscsi1
Enabled hdisk1 fscsi1
Enabled hdisk2 fscsi1
```

7.3 How Can I Set a Load Balancing Mode for a Disk?

Question

How can I set a load balancing mode for a disk?

Answer

The ODM software supports three load balancing modes: fail_over, round_robin, and shortest_queue. The default mode is fail_over.

The fail_over mode is an optimal path mode. You can set different priorities for paths. Application servers select the path with the highest priority to deliver I/Os.
NOTE:
- The path priority depends on the value of the priority parameter. A smaller value of the priority parameter indicates a higher priority. The value of the priority parameter ranges from 1 to 255. If the value of the priority parameter is set to 1, the path priority is the highest. If the value of the priority parameter is set to 255, the path priority is the lowest.
- By default, the value of the priority parameter is 1.
For the round_robin mode, when an application server delivers I/Os to a storage system for the first time, the I/Os are transferred through path 1. When the I/Os are delivered for the secondary time, path 2 is used. The paths that will be used are deduced by analogy. Paths are used in turn to ensure that each path is fully utilized.
For the shortest_queue mode, when an application server delivers I/Os to a storage system, the I/Os are transferred through a path that currently has the fewest I/Os. If the transfer speed of a path is slow, the other paths whose transfer speed is quick can be fully utilized.
NOTE:

This mode can only be set in some operating systems that adopt the latest technologies. For details, see the official website of IBM.

To set a load balancing mode for a disk, do as follows:

Query the disk properties.

Run the lsattr -El hdiskx command to query the disk properties (hdisk1 as an example). The value of the algorithm field indicates the path selection algorithm. The default value is fail_over.

-bash-3.00# lsattr -El hdisk1
PCM             PCM/friend/MPIOpcm               Path Control Module              False
PR_key_value    none                             Persistant Reserve Key Value     True
algorithm       fail_over                        Algorithm                        True
clr_q           no                               Device CLEARS its Queue on error True
dist_err_pcnt   0                                Distributed Error Percentage     True
dist_tw_width   50                               Distributed Error Sample Time    True
hcheck_cmd      test_unit_rdy                    Health Check Command             True
hcheck_interval 30                               Health Check Interval            True
hcheck_mode     nonactive                        Health Check Mode                True
location                                         Location Label                   True
lun_id          0x1000000000000                  Logical Unit Number ID           False
lun_reset_spt   yes                              LUN Level Reset                  True
max_transfer    0x40000                          Maximum TRANSFER Size            True
node_name       0x21000022a1073fdd               FC Node Name                     False
pvid            00007e1d562bd42d0000000000000000 Physical volume identifier       False
q_err           yes                              Use QERR bit                     True
q_type          simple                           Queuing TYPE                     True
queue_depth     32                               Queue DEPTH                      True
reassign_to     120                              REASSIGN time out value          True
reserve_policy  no_reserve                       Reserve Policy                   True
rw_timeout      30                               READ/WRITE time out value        True
scsi_id         0x10400                          SCSI ID                          False
start_timeout   60                               START unit time out value        True
ww_name         0x20190022a1073fdd               FC World Wide Name               False

Set a load balancing mode for the disk.
- Run the chdev -l hdisk1 -a algorithm=round_robin command to set the load balancing mode of the disk to round_robin.
- Run the chdev -l hdisk1 -a algorithm=fail_over command to set the load balancing mode of the disk to fail_over.
- Run the chdev -l hdisk1 -a algorithm=shortest_queue command to set the load balancing mode of the disk to shortest_queue.

Optional: If the load balancing mode is fail_over, set path priorities.

Query path priorities.

Run the lspath -F"name:parent:connection:path_id:status" command to query path priorities. The command output is as follows:

-bash-3.00# lspath -F"name:parent:connection:path_id:status"
hdisk0:scsi1:8,0:0:Enabled
hdisk1:fscsi0:20080022a10bb2d5,1000000000000:0:Enabled
hdisk2:fscsi0:20080022a10bb2d5,2000000000000:0:Enabled
hdisk3:fscsi0:20080022a10bb2d5,3000000000000:0:Enabled
hdisk4:fscsi0:20080022a10bb2d5,4000000000000:0:Enabled
hdisk5:fscsi0:20080022a10bb2d5,5000000000000:0:Enabled
hdisk6:fscsi0:20080022a10bb2d5,6000000000000:0:Enabled
hdisk7:fscsi0:20080022a10bb2d5,7000000000000:0:Enabled
hdisk8:fscsi0:20080022a10bb2d5,8000000000000:0:Enabled
hdisk1:fscsi0:20090022a10bb2d5,1000000000000:1:Enabled
hdisk2:fscsi0:20090022a10bb2d5,2000000000000:1:Enabled
hdisk3:fscsi0:20090022a10bb2d5,3000000000000:1:Enabled
hdisk4:fscsi0:20090022a10bb2d5,4000000000000:1:Enabled
hdisk5:fscsi0:20090022a10bb2d5,5000000000000:1:Enabled
hdisk6:fscsi0:20090022a10bb2d5,6000000000000:1:Enabled
hdisk7:fscsi0:20090022a10bb2d5,7000000000000:1:Enabled
hdisk8:fscsi0:20090022a10bb2d5,8000000000000:1:Enabled
hdisk1:fscsi0:20180022a10bb2d5,1000000000000:2:Enabled
hdisk2:fscsi0:20180022a10bb2d5,2000000000000:2:Enabled
hdisk3:fscsi0:20180022a10bb2d5,3000000000000:2:Enabled
hdisk4:fscsi0:20180022a10bb2d5,4000000000000:2:Enabled
hdisk5:fscsi0:20180022a10bb2d5,5000000000000:2:Enabled
hdisk6:fscsi0:20180022a10bb2d5,6000000000000:2:Enabled
hdisk7:fscsi0:20180022a10bb2d5,7000000000000:2:Enabled
hdisk8:fscsi0:20180022a10bb2d5,8000000000000:2:Enabled
hdisk1:fscsi0:20190022a10bb2d5,1000000000000:3:Enabled
hdisk2:fscsi0:20190022a10bb2d5,2000000000000:3:Enabled
hdisk3:fscsi0:20190022a10bb2d5,3000000000000:3:Enabled
hdisk4:fscsi0:20190022a10bb2d5,4000000000000:3:Enabled
hdisk5:fscsi0:20190022a10bb2d5,5000000000000:3:Enabled
hdisk6:fscsi0:20190022a10bb2d5,6000000000000:3:Enabled
hdisk7:fscsi0:20190022a10bb2d5,7000000000000:3:Enabled
hdisk8:fscsi0:20190022a10bb2d5,8000000000000:3:Enabled

Run the lspath -l hdiskx -a priority -F value -p fscsix -w connection command to check the priority of a specified path.
```
-bash-3.00# lspath -l hdisk1 -a priority -F value -p fscsi0 -w 20080022a10bb2d5,1000000000000
1
```

Set path priorities.
Run the chpath -l hdiskx -a priority=x -p fscsi0 -w connection command to modify the priority of a specified path. The priority value ranges from 1 to 255 and 1 indicates the highest priority.
```
-bash-3.00# chpath -l hdisk1 -a priority=255 -p fscsi0 -w 20080022a10bb2d5,1000000000000
path Changed
-bash-3.00# lspath -l hdisk1 -a priority -F value -p fscsi0 -w 20080022a10bb2d5,1000000000000
255
```

7.4 How Can I Configuring the ReserveCheck and ReserveBreak Functions?

Question

How can I configuring the ReserveCheck and ReserveBreak functions?

Answer

The odmget -q name = xxxx CuDv and smit hacmp commands are used to configure the ReserveCheck and ReserveBreak functions. The name indicates device name. If one node fails, another node automatically invokes ReserveCheck and ReserveBreak to take over the resources on the failed node.

NOTE:

Before configuring these functions, ensure that the HACMP software has been installed.

Perform the following steps to configure the ReserveCheck and ReserveBreak functions:

Run odmget -q name = xxxx CuDv to obtain the disk type.
Run smit hacmp.
On the HACMP for AIX configuration page, choose Extended Configuration > Extended Resource Configuration > HACMP Extended Resources Configuration > Configure Custom Disk Methods > Add Custom DiskMethods.

On the following output, enter the disk type in New Disk Type.

[Entry Fields]
* New Disk Type []
* Method to identify ghost disks
[SCSI3] +
* Method to determine if a reserve is held [/usr/sbin/
ReserveCheck] +
* Method to break a reserve [/usr/sbin/
ReserveBreak] +
Break reserves in parallel
false +
* Method to make the disk available [MKDEV]

Press Enter to finish configuring the ReserveCheck and ReserveBreak functions.

For example, configure the ReserveCheck and ReserveBreak functions for hdisk2 (the disk type is disk/fcp/HWS5300).

# odmget -q 'name = hdisk2' CuDv
CuDv:
name = "hdisk2"
status = 1
chgstatus = 0
ddins = "scsidisk"
location = "1D-08-02"
parent = "fscsi1"
connwhere = "W_0"
PdDvLn = "disk/fcp/HWS5300"

# smit hacmp
[Entry Fields]
* New Disk Type 								[disk/fcp/HWS5300]
* Method to identify ghost disks					[SCSI3]+
* Method to determine if a reserve is held 			[/usr/sbin/ReserveCheck] +
* Method to break a reserve 						[/usr/sbin/ReserveBreak] +
Break reserves in parallel						false +
* Method to make the disk available 				[MKDEV]

7.5 How Do I Use UltraPath for AIX to Replace AIX ODM for MPIO in SAN Boot Scenarios?

Question

How do I use UltraPath for AIX to replace AIX ODM for MPIO in SAN boot scenarios?

Answer

Uninstall AIX ODM for MPIO following 5 Uninstalling AIX ODM for MPIO and restart the AIX system.

After the AIX system is started, run bootlist -m normal -o hdiskXX to set the boot list and run shutdown -Fr to restart the AIX system. You can run lslv -m hd5 to query hdiskXX. For example,

bash-3.00# lslv -m hd5
hd5:N/A
LP    PP1  PV1               PP2  PV2               PP3  PV3
0001  0001 hdisk9            
bash-3.00# bootlist -m normal -o hdisk9
hdisk9 blv=hd5
bash-3.00# 
bash-3.00# shutdown -Fr

SHUTDOWN PROGRAM
Tue Apr 11 11:48:02 CDT 2017

After the AIX system is started, run lsdev -Cc disk to check the residual defined and N/A disks. For example:

bash-3.00# lsdev -Cc disk
lsdev: 0514-521 Cannot find information in the predefined device
        configuration database for the customized device hdisk8.
hdisk0  Available          Virtual SCSI Disk Drive
hdisk1  Available C4-T1-01 Other FC SCSI Disk Drive
hdisk2  Available C4-T1-01 Other FC SCSI Disk Drive
hdisk3  Available C4-T1-01 Other FC SCSI Disk Drive
hdisk4  Available C4-T1-01 Other FC SCSI Disk Drive
hdisk5  Available C4-T1-01 Other FC SCSI Disk Drive
hdisk6  Available C4-T1-01 Other FC SCSI Disk Drive
hdisk7  Available C4-T1-01 Other FC SCSI Disk Drive
hdisk8  Defined   C4-T1-01 N/A
hdisk9  Available C4-T1-01 Other FC SCSI Disk Drive

hdisk8 is the residual disk information.

Delete the residual disk information and run the following commands for the disk:

odmdelete -o CuDv -q name=hdiskX 
odmdelete -o CuAt -q name=hdiskX
odmdelete -o CuPath -q name=hdiskX
odmdelete -o CuPathAt -q name=hdiskX
odmdelete -o CuVPD -q name=hdiskX
odmdelete -o CuDvDr -q value3=hdiskX

Run bosboot -a to update the boot image. For example:

bash-3.00# odmdelete -o CuDv -q name=hdisk8
0518-307 odmdelete: 1 objects deleted.
bash-3.00# odmdelete -o CuAt -q name=hdisk8
0518-307 odmdelete: 6 objects deleted.
bash-3.00# odmdelete -o CuPath -q name=hdisk8
0518-307 odmdelete: 2 objects deleted.
bash-3.00# odmdelete -o CuPathAt -q name=hdisk8
0518-307 odmdelete: 4 objects deleted.
bash-3.00# odmdelete -o CuVPD -q name=hdisk8
0518-307 odmdelete: 1 objects deleted.
bash-3.00# odmdelete -o CuDvDr -q value3=hdisk8
0518-307 odmdelete: 1 objects deleted.
bash-3.00# bosboot -a

bosboot: Boot image is 53276 512 byte blocks.

Run sh install.sh to install UltraPath. Restart the AIX system after the installation.
If the AIX system fails to be started, go to the single user mode and run lsdev -Cc disk to check whether a defined disk exists. Run rmdev -dl hdiskXX to remove the defined disk. Execute bootlist -m normal -o hdiskXX to specify a boot device and run bosboot -a. Restart the AIX system.
For details about how to go to the single user mode, refer to http://www-01.ibm.com/support/docview.wss?uid=isg3T1011796.

http://support.huawei.com/enterprise/en/doc/EDOC1000087221?currentPartNo=j00a&togo=content