Power

Hi there,

Reference case# TS010294786

Due to the "Virtual Storage Redundancy with dual VIOS Configuration"
URL: https://community.ibm.com/community/user/power/blogs/robert-kovacs1/2020/07/22/virtual-storage-redundancy-with-dual-vios-configur

In  the figure 4 and 7 from above link there are similar to current customer environment.

Could you please help to evaluate the capability if customer face some various situation issue.

1. Retrieve some of necessary parameters were current settings from snap data in both VIOS and VIOS level. (may be helpful)

Reference the current solution was designed for AIX production servers Env to preventing data lose by mirroring all VGs
(such as rootvg and other VG) by using the hdisks from the difference storage sources with NPIV features (From 2 DS8000 boxes).

Herewith the parameters setting on VIOS and VIOC.

A. In VIOS level, current settings on VIO Server : p9viosrt1-p1

   - fscsi device on VIOS.
       - dyntrk yes
       - fc_err_recov fast_fail

B. In VIOC level, current settings on VIOC : eaiqm1

   - Hdisk device in VIO Client.

      - algorithm shortest_queue
      - reserve_policy no_reserve
      - hcheck_interval 60
      - hcheck_mode nonactive
      - queue_depth 256

   - fscsi devices in client:

      - dyntrk yes
       - fc_err_recov fast_fail


- Example MPIO path access for hdisk0 and hdisk1

   hdisk0:
      name path_id status path_status parent connection
      hdisk0 0 Enabled Sel fscsi0 50050763080b10d6,4036406000000000
      hdisk0 1 Enabled Sel fscsi1 50050763081b10d6,4036406000000000
      hdisk0 2 Enabled Sel fscsi2 50050763081bd0d6,4036406000000000
      hdisk0 3 Enabled Sel fscsi3 50050763080bd0d6,4036406000000000

   hdisk1:
     name path_id status path_status parent connection
     hdisk1 4 Enabled Sel fscsi0 50050763081390d8,4036406000000000
     hdisk1 5 Enabled Sel fscsi1 50050763081390d8,4036406000000000
    hdisk1 6 Enabled Sel fscsi2 50050763081b50d8,4036406000000000
    hdisk1 7 Enabled Sel fscsi3 50050763081b50d8,4036406000000000


2. Could you please help us to evaluate the capability if customer face some various situation issue in the future.

Is it possible all of those parameters in VIOS and VIOC settings are applicable/appropriate and it could help to preventive for each issue scenarios below?


Q1. When 1 out of 2 DS8000 failure (1 DS8000 goes down , i.e power outage) issue?
Means, it should be there is not impact to AIX server were having a disk from remaining 1 DS8000 were mirrored and able running as normal.
Please help to evaluate and feedback your advice.


Q2. When 1 out of 2 DS8000 goes slow down or the I/O response is degrade/delay (DS8000 not failure, but slow response when some of DS8000 controller have HW failure) issue?
Means, it should be there is some impact to AIX server such as system goes slow down response, but not goes to hang state or unable login.
please help to evaluate and feedback your advice.

In addition, refer the old case #TS005111653 workaround fix by unmirror all VGs, Is it the right solution? or there any else should be consider.

Q3. When on VIOS, VIOC and DS800 have some of HW failure issue for some of component i.e. HBA card, fiber adapter fail, DS8000 controller failure...etc..?
Means, in VIOC and VIOS still using the available paths access to reach to disk in both DS8000 storage and it could running as normal.
please help to evaluate and feedback your advice.

Regards,
Charin Kumjudpai (AIX local support)