Sponsored Content
Full Discussion: DISK and MPIO
Operating Systems AIX DISK and MPIO Post 303033199 by bakunin on Monday 1st of April 2019 11:11:09 AM
Old 04-01-2019
Quote:
Originally Posted by Phat
Yes I know Smilie but can we pickup whatever 1 disk and make it use AIX PCM or vice versa? Just want to know the possibility.
Well, maybe i didn't phrase that clearly enough: that depends. It depends on a lot of things i don't know because i do not know your environment and - with all due respect - you don't seem to know either, so you can't tell me what i would need to know to explain it to you.

Quote:
Originally Posted by Phat
But I wonder why how they can have the 2 different drivers. I detect hdisk18 belongs to IBMSVC which is using AIX PCM whereas the others are using SDDPCM.
SDDPCM and MPIO are two different multipathing drivers, both for AIX. SDDPCM is a bit older than MPIO. You cannot just switch between them because you will need to meet certain requirements to use either of them. But this would be better discussed with someone knowing your complete setup: fabric, storage boxes, zoning, FC-switches, virtualiser, .... I would be of limited help in this capacity (and everybody else here too) because we will not have the intimate knowledge that would be required. My suggestion: go to the storage people and ask one of them to explain storage basics and their application in your specific surroundings to you.

I am terribly sorry but this is something i cannot tell you over the net with just a few forum posts. The matter is just too complex (and the single pieces too interdependent) for that.

Quote:
Originally Posted by Phat
So what is IBMSVC? can we move disk from IBMSVC to 2017DS8K? I can see that all the LUNs comes from 1 source storage and visible to all 4 FCs adapters.
IBM SAN Volume Controller is a storage virtualiser. It is similar in scope to i.e. EMCs VPlex product. The DS8K is a (high-end) storage box itself, which is probably managed (or maybe not - i don't know your environment) by the SVC. In principle the SVC sits between (all) the storage boxes on one side and (all) the systems using their LUNs on the other side. So, the answer to your question wether you can move the LUN from one to the other: maybe. It depends on wether the DS8k is configured to let the SVC manage the LUNs it exports or not and, again, it depends on the layout of your FC fabric and it depends on your FC switches and how they are connected and zoned and, .... You see the problem?

I hope this helps.

bakunin
 

10 More Discussions You Might Find Interesting

1. UNIX for Advanced & Expert Users

AIX MPIO and EMC

We are looking at running MPIO for it's redundancy and load balancing benefits. Does anyone know what pieces of software or modules are needed on the VIO server to get load balancing to work. Remember we are using EMC's DMX3500 storage system. We no longer want to use Powerpath. :rolleyes: ... (2 Replies)
Discussion started by: vxg0wa3
2 Replies

2. AIX

AIX 5.3 MPIO vs RDAC on DS4000 range

Hi I would like to ask what the benefits are of changing from RDAC to MPIO when connecting to a DS4000 on AIX 5.3? I have heard that IBM MPIO "might" support more than 1 active path to a LUN when connecting to a DS4800 through more than 1 host connection on the same AIX client. I understand that... (8 Replies)
Discussion started by: kimyo
8 Replies

3. AIX

AIX native MPIO

Hi folks, does anybody have a link to a documentation how to implement native MPIO on AIX? We are using EMC PowerPath and Datacore SanSymphony/Cambex for this so far and I wasn't able to find a good description on that topic. All I know so far is that mkpath, chpath and lspath are used to... (3 Replies)
Discussion started by: zaxxon
3 Replies

4. AIX

SDD SDDPCM MPIO lspath Jargon

Can anyone recommend me some reading material surrounding how AIX handles LUNs: - with and without MPIO installed - with and without SDD or SDDPCM installed Where does lspath sit in all of this (MPIO layer?). Can a system be built with just MPIO software? Is MPIO software even needed? I guess... (0 Replies)
Discussion started by: apra143
0 Replies

5. AIX

MPIO Driver

On a particular LPAR, I was running AIX 5.3 TL 3. On Monday I did an update of the LPAR to 5.3 TL 9 SP2. The install was smooth, but then I ran into a problem. The MPIO driver does not work with LSI's StoreAge (SVM4). I did some looking, and looks like 5.3 TL3 = IBM.MPIO 5.3.0.30 5.3... (0 Replies)
Discussion started by: clking
0 Replies

6. AIX

MPIO reliability

Hi, we have a vew boxes using MPIO and they are connected to some virtualization software managing some disk subsystems, offering volumes to the AIX boxes. Sometimes when a cable has been plugged out for a test or when a real problem occurs, using lspath to show the state of the paths shows... (8 Replies)
Discussion started by: zaxxon
8 Replies

7. AIX

MPIO - list of supported arrays

Hello, we are planning to deploy some of our databases on AIX/LPAR based servers (we didn't bought it yet ...). IBM's engineers says that if we want to boot them from SAN the hardware array has to be compatible with MPIO but they don't want to deliver any document with list of arrays supported... (5 Replies)
Discussion started by: andy55
5 Replies

8. AIX

Upgrading from native MPIO to SDDPCM AIX 6.1

We have AIX 6.1 system attached to SAN disks (DS4700 and DS8100) thru SVC. Initially when the system was I forgot to install sddpcm drivers. and wanted to know how can i go with the installation of the sddpcm drivers. My understandin going thru the manual ... (3 Replies)
Discussion started by: mk8570
3 Replies

9. AIX

Failed mpio path on AIX5.3

I have found failed mpio path on AIX and enabled that failed path as below failed path on AIX bash-3.2# lspath -l hdisk10 Enabled hdisk10 fscsi0 Enabled hdisk10 fscsi0 Failed hdisk10 fscsi3 Enabled hdisk10 fscsi3 Enabled failed path as below chpath -l hdisk10 -p fscsi3 -s enable... (13 Replies)
Discussion started by: murali969
13 Replies

10. AIX

Need Help with SDD / SDDPCM / MPIO

This is getting very confusing for me, and appreciate if someone can help. Platform: Power VM ( Virtual I/O Server) ioslevel 2.1.3.10-FP23 # oslevel -s 6100-05-00-0000 Storage: IBM DS4300 Two HBAs - Dual Port Fibre Adapter Channels Each card has two ports , so a total of 4 ports going... (3 Replies)
Discussion started by: filosophizer
3 Replies
cfgadm_fp(1M)															     cfgadm_fp(1M)

NAME
cfgadm_fp - driver specific commands for cfgadm SYNOPSIS
/usr/sbin/cfgadm [-f] [-n | -y] [-v] [-o hardware_options] -c function ap_id [ap_id] /usr/sbin/cfgadm [-v] [-a] [-s listing_options] [-o hardware_options] [-l [ap_id]] /usr/sbin/cfgadm [-v] [-o hardware_options] -h [ap_id] The fp port driver plug-in /usr/lib/cfgadm/fp.so.1 provides the functionality for Fibre Channel Fabric device node management through cfgadm(1M). cfgadm operates on attachment points. Attachment points are locations in the system where hardware resources can be dynamically reconfigured. Refer to cfgadm(1M) for additional details on attachment points. For Fibre Channel Fabric device node management, each fp port node is represented by an attachment point in the device tree. In addition, each Fibre Channel device is represented by a dynamic attachment point. Attachment points are named through ap_ids. Two types of ap_ids are defined: logical and physical. The physical ap_id is based on the physical pathname. The logical ap_id is a shorter, more user-friendly name. For fp port nodes, the logical ap_id is the corresponding disk controller number. For example, c0 is a typical logical ap_id. Fibre Channel devices are named with a port World Wide Name (WWN). If a disk device is connected to controller c0, its ap_id can be: c0::50020f2300006077 where 50020f2300006077 identifies the port WWN of a specific Fibre Channel device. Each device on the Fibre Channel private loop port, Fabric port or public loop port is probed and made available to Solaris by default. Devices connected to the Fibre Channel Fabric port or public loop port can be made unavailable to Solaris by initiating an application or an end user operation. The operation is similar to the hot unplugging of devices by way of management user interfaces. Applications or users can use the /usr/lib/cfgadm/fp.so.1 library to enable libcfgadm to provide interfaces to accomplish this task. The list of currently connected Fabric devices is generated in the form of the attachment point. A simple listing of attachment points in the system includes attachment points at fp port nodes but not Fibre Channel devices. The follow- ing example uses the -a flag to the list option (-l) to list Fibre Channel devices: # cfgadm -l Ap_Id Type Receptacle Occupant Condition c0 fc-fabric connected configured unknown c1 fc-private connected configured unknown c2 fc connected unconfigured unknown sysctrl0:slot0 cpu/mem connected configured ok sysctrl0:slot1 sbus-upa connected configured ok The following example lists Fibre Channel devices connected to fp ports. # cfgadm -al Ap_Id Type Receptacle Occupant Condition c0 fc-fabric connected configured unknown c0::50020f2300006077 disk connected configured unknown c0::50020f23000063a9 disk connected configured unknown c0::50020f2300005f24 disk connected configured unknown c0::50020f2300006107 disk connected configured unknown c1 fc-private connected configured unknown c1::220000203708b69c disk connected configured unknown c1::220000203708ba7d disk connected configured unknown c1::220000203708b8d4 disk connected configured unknown c1::220000203708b9b2 disk connected configured unknown c2 fc connected unconfigured unknown sysctrl0:slot0 cpu/mem connected configured ok sysctrl0:slot1 sbus-upa connected configured ok In this example, the fc-fabric type of ap_id c0 indicates that the fp port is connected to Fabric. For an fp port with Fabric related type such as fc-fabric and fc-public, device node creation happens by default at the boot time and can be managed by the cfgadm configure and unconfigure operations. The fc-private type of ap_id c1 indicates that fp port is connected to private-loop and device node creation hap- pens by default as well. The fc type of ap_id c2 indicates that nothing is attached to fp port c2. The Type field of a Fibre Channel device ap_id shows the SCSI device type of LUN 0 in the device. A Fibre Channel device with multiple FCP SCSI LUNs is configured into Solaris and each FCP SCSI LUN is available as a Solaris device. Sup- pose that ap_ids c0::50020f2300006077 and c0::50020f23000063a9 represent Fibre Channel devices with multiple FCP SCSI LUNs. The following example shows how to list ap_ids with FCP SCSI LUN information: # cfgadm -al -o show_SCSI_LUN Ap_Id Type Receptacle Occupant Condition c0 fc-fabric connected configured unknown c0::50020f2300006077,0 disk connected configured unknown c0::50020f2300006077,1 disk connected configured unknown c0::50020f2300006077,2 disk connected configured unknown c0::50020f2300006077,3 disk connected configured unknown c0::50020f23000063a9,0 disk connected configured unknown c0::50020f23000063a9,1 disk connected configured unknown c0::50020f23000063a9,2 disk connected configured unknown c0::50020f23000063a9,3 disk connected configured unknown c0::50020f2300005f24,0 disk connected unconfigured unknown c0::50020f2300005f24,1 disk connected unconfigured unknown c0::50020f2300006107,0 disk connected unconfigured unknown c0::50020f2300006107,1 disk connected unconfigured unknown c1 fc-private connected configured unknown c1::220000203708b69c,0 disk connected configured unknown c1::220000203708ba7d,0 disk connected configured unknown c1::220000203708b8d4,0 disk connected configured unknown c1::220000203708b9b2,0 disk connected configured unknown c2 fc connected unconfigured unknown In this example, the ap_id c0::50020f2300006077,0 identifies the FCP SCSI LUN 0 of the Fibre Channel device which is represented by port WWN 50020f2300006077. The Fibre Channel device is reported to have 4 FCP SCSI LUNs and they are all configured. 4 FCP SCSI LUN level ap_ids associated with port WWN 50020f2300006077 are listed. The listing also displays FCP SCSI LUNs for unconfigured Fibre Channel devices. The Fibre Channel device represented by c0::50020f2300005f24 is reported to have two FCP SCSI LUNs. The configure operation on c0::50020f2300005f24 creates two Solaris devices. The Type field of FCP SCSI LUN level ap_ids show the SCSI device type of each LUN. When a Fibre Channel device has different device type LUNs, the Type field reflects that. The receptacle and occupant state for attachment points at the fp port have the following meanings: configured One or more devices configured on the fp port connected fp port active disconnected fp port quiesced (IO activity is suspended) empty Not applicable unconfigured No devices configured on the fp port The state for individual Fibre Channel devices on an fp port: configured Device is configured into Solaris and is available for use connected fp port to which the device is connected to is active disconnected fp port to which the device is attached is quiesced unconfigured Device is available to be configured The condition field for attachment points at the fp port has the following meanings: failed An error condition has prevented the fp port from being able to detect the presence or type of a Fibre Channel con- nection. The condition field for individual Fibre Channel devices on an fp port has the following meanings: failed An error is encountered while probing a device on Fabric. failing A device was configured on a host and its state as seen by Solaris appears to be normal (i.e., online) but it is either not currently present or visible in the fabric or its presence could not be verified due to an error condi- tion on the local port through which the device was configured. unusable A device has been configured on the host, but is currently offline or failed. The unknown condition indicates that probing a device on Fabric completed without an error and the device state within Solaris host is nor- mal if the device was configured previously. The internal condition of the device cannot be guaranteed. cfgadm defines several types of operations in addition to listing (-l). These operations include invoking configuration state changes and obtaining configuration administration help messages (-h). The following options are supported: -c function The following generic commands are defined for the fp-transport-specific library: For Fibre Channel device attachment points on the fc-fabric type fp port attachment point, the following configuration state change operations are supported: configure Configure a connected Fibre Channel Fabric device to a host. When a Fibre Channel device is listed as an unknown type in the output of the list operation the device might not be configurable. No attempt is made to configure devices with unknown types. The force option (-f) can be used to force the fp port driver plug-in to make an attempt to configure any devices. Any errors in the process are reported. See NOTES. unconfigure Unconfigure a Fibre Channel Fabric device from a host. This device stays unconfigured until the next reboot or Solaris Dynamic Reconfiguration on the controller that the device is connected, at which time all fabric devices are automati- cally enumerated. The default behavior may be changed through the use of the "manual_configuration_only" property in the fp.conf file. If the property is set, the device remains unconfigured after reboot. Refer to fp(7D) for additional details on the "manual_configuration_only" property. For Fibre Channel private loop devices, the configure command returns success without doing any operation. The unconfigure command is not supported on the private loop devices. The private loop devices are configured by Solaris Fibre Channel drivers by default and are not managed through end user- or application-initiated operations. -f Force the configure change state operation to occur irrespective of the condition or type. Refer to the above description of the con- figure change state operation. -h ap_id Obtain fp--transport-specific help. Specify any fp attachment point. -o hardware_options The following hardware options are supported. show_SCSI_LUN Lists ap_ids associated with each FCP SCSI LUN for discovered Fibre Channel devices when specified with the list option -al. Refer to the previously mentioned description and example of FCP SCSI LUN level listing. Device node creation is not supported on the FCP SCSI LUN level. See NOTES. All Fibre Channel devices are available to Solaris by default. Enabling only a subset of Fabric devices available to Solaris by default can be accomplished by setting the property "manual_configuration_only" in /kernel/drv/fp.conf file. When "manual_configuration_only" in fp.conf is set, all Fabric devices are not available to Solaris unless an application or an end user had previously requested the device be configured into Solaris. The configure state-change command makes the device available to Solaris. After a successful config- ure operation on a Fabric device, the associated links are added to the /dev namespace. The unconfigure state-change command makes a device unavailable to Solaris. When a Fibre Channel Fabric device is configured successfully to a host using the -c configure operation, its physical ap_id is stored in a repository. When a Fibre Channel Fabric device is unconfigured using the -c unconfigure operation, its physical ap_id is deleted from the same repository. All fabric devices are automatically enumerated by default and the repository is used only if the fp.conf "manual_configuration_only" property is set. Refer to fp(7D) for additional details on the "manual_configuration_only" property. You can specify the following commands with the -c option to control the update behavior of the repository: force_update For configure, the attachment point is unconditionally added to the repository; for unconfigure, the attachment point is unconditionally deleted. no_update No update is made to the repository regardless of the operation. These options should not be used for normal configure and unconfigure operations. See WARNINGS. When a Fibre Channel device has multiple FCP SCSI LUNs configured and any Solaris device associated with its FCP SCSI LUN is in the unusable condition, the whole Fibre Channel device is reported as unusable. The following option with the -c unconfigure command removes only Solaris devices with the unusable condition for a Fibre Channel device. unusable_SCSI_LUN For unconfigure operation, any offlined device nodes for a target device is removed. -s listing_options Refer to cfgadm(1M) for usage information. -t ap_id No test commands are available at present. -x hardware_function No hardware specific functions are available at present. All other options have the same meaning as defined in the cfgadm(1M) man page. Example 1: Unconfiguring a Disk The following command unconfigures a disk: # cfgadm -c unconfigure c0::210000203708b606 Example 2: Unconfigure all the Configured Disks under Single Attachment Point The following command unconfigures all configured disks under the attachment point c0. # cfgadm -c unconfigure c0 Example 3: Configuring a Disk The following command configures a disk: # cfgadm -c configure c0::210000203708b606 Example 4: Configure all the Unconfigured Disks under Single Attachment Point The following command configures all unconfigured disks under the attachment point c0. # cfgadm -c configure c0 Example 5: Removing the Fibre Channel Fabric Device Attachment Point from Repository The following command unconditionally removes the fibre channel fabric device attachment point from the Fabric device repository. # cfgadm -c unconfigure -o force_update c0::210000203708b606 Example 6: Removing Offlined Solaris Device Nodes for a Target Device The following command removes offlined Solaris device nodes for a target device: # cfgadm -c unconfigure -o unusable_SCSI_LUN c0::210000203708b606 FILES
/usr/lib/cfgadm/fp.so.1 Hardware-specific library for Fibre Channel Fabric device node management. /etc/cfg/fp/fabric_WWN_map Repository of physical ap_ids of Fabric devices currently configured. It is used only to reconfigure those Fabric devices at boot time. This repository is only used when the "manual_configuration_only" /kernel/drv/fp.conf file is set. /etc/rcS.d/fdevattach Reconfigures Fabric device(s) of which physical ap_id is listed in /etc/cfg/fp/fabric_WWN_map on boot time. See attributes(5) for descriptions of the following attributes: +-----------------------------+-----------------------------+ | ATTRIBUTE TYPE | ATTRIBUTE VALUE | +-----------------------------+-----------------------------+ |Availability |SUNWcfpl, SUNWcfplx | +-----------------------------+-----------------------------+ svcs(1), cfgadm(1M), svcadm(1M), config_admin(3CFGADM), libcfgadm(3LIB), attributes(5), smf(5), fp(7D) WARNINGS
Do not use hardware-specific options for the repository update under normal configure/unconfigure operations. The hardware-specific options are expected to be used when the node creation of a Fabric device fails at boot time and the error condition is considered to be permanent. The unconfigure command with force_update hardware-specific option unconditionally removes the attachment point of a failing Fabric device from the repository. For devices with unknown or no SCSI device type (for example, a Fibre Channel Host Bus Adapter), the configure operation might not be applicable. The configure and unconfigure commands operate on the Fibre Channel device level which is represented by port WWN ap_id. If a Fibre Channel device has multiple FCP SCSI LUNs configured, the configure command on the associated port WWN ap_id results in creating a Solaris device for each FCP SCSI LUN. The unconfigure command removes all Solaris devices associated with the port WWN ap_id. The FCP SCSI LUN level ap_id is not valid for the configure and unconfigure commands. The deprecated show_FCP_dev option has been replaced by the new show_SCSI_LUN option, and the deprecated unusable_FCP_dev option has been replaced by the new unusable_SCSI_LUN option. The cfgadm_fp service is managed by the service management facility, smf(5), under the service identifier: svc:/system/device/fc-fabric:default Administrative actions on this service, such as enabling, disabling, or requesting restart, can be performed using svcadm(1M). The ser- vice's status can be queried using the svcs(1) command. No administrative actions on this service are required for Fabric device configuration once this service is started on boot time. 2 Nov 2004 cfgadm_fp(1M)
All times are GMT -4. The time now is 07:38 AM.
Unix & Linux Forums Content Copyright 1993-2022. All Rights Reserved.
Privacy Policy