HMC

Guys,

Does anyone use such functionality in production environment where it is regularly tested?

We want to get rid of NIMs and use "installios" HMC command which potentially should restore VIOS backup from a tar image. The problem is this is not a bullet proof solution. It seems it may work, or it may not. The investigation is always difficult because it would require PESH password, few PMRs. From what I see, it can't be used as a resilient scenario. 

What is your experience, mainly IBMi shops? What is your VIOS resiliency strategy? I am not asking AIX community. For this group, the NIM server is a must. 

Hi,

Here, I have:

• scheduled "IO Configuration" for each VIOS, on 2 different HMC
• Regularly refreshed sysplans
• VIOS installation image at hand (customized for non-IBM storage)

The idea is to reinstall the VIOS, then use rstviosbk. However, I didn't test this recovery plan as VIOS are very unlikely to fail completely ;)

I used a very similar process to upgrade from PowerVM 3 to 4, as the automated upgrade won't work with our non-IBM storage. The only difference is that configuration was kept, I didn't need to restore it.

I'm not aware of PESH password nor PMRs, what is your hardware configuration?

Guys,

Does anyone use such functionality in production environment where it is regularly tested?

We want to get rid of NIMs and use "installios" HMC command which potentially should restore VIOS backup from a tar image. The problem is this is not a bullet proof solution. It seems it may work, or it may not. The investigation is always difficult because it would require PESH password, few PMRs. From what I see, it can't be used as a resilient scenario. 

What is your experience, mainly IBMi shops? What is your VIOS resiliency strategy? I am not asking AIX community. For this group, the NIM server is a must. 

The idea is to reinstall the VIOS, then use rstviosbk. However, I didn't test this recovery plan as VIOS are very unlikely to fail completely ;)

 :) well this is not an excuse not to have the resiliency plan. I can give you two scenarios where we had to recover VIOS image.

a) new version of VIOS did not work properly. It was discovered immediately after VIO upgrade (SEA did not work - VIOS bug), and we had to recover from the backup

b) External storage failure - complete failure of full redundant enterprise external storage, with mirroring and all features - firmware bug. 

Restore of VIOS config, is not really option for us. We have dozen of VIOSes, on different POWER models, with different HW configuration. We do not have always the same VIOS release everywhere. It would be extremely difficult to restore VIOS image and then apply each specific fix on each for the particular VIOS. The RTO would be extremely prolonged (which is not an option in enterprise configuration).  

We have HW based HMCs everywhere, and dealing with PESH password is a must if you are dealing with misbehaving of  HMC software. 

https://www.ibm.com/support/pages/hmc-installios-cleanup

Hi,

Here, I have:

• scheduled "IO Configuration" for each VIOS, on 2 different HMC
• Regularly refreshed sysplans
• VIOS installation image at hand (customized for non-IBM storage)

The idea is to reinstall the VIOS, then use rstviosbk. However, I didn't test this recovery plan as VIOS are very unlikely to fail completely ;)

I used a very similar process to upgrade from PowerVM 3 to 4, as the automated upgrade won't work with our non-IBM storage. The only difference is that configuration was kept, I didn't need to restore it.

I'm not aware of PESH password nor PMRs, what is your hardware configuration?

Guys,

Does anyone use such functionality in production environment where it is regularly tested?

We want to get rid of NIMs and use "installios" HMC command which potentially should restore VIOS backup from a tar image. The problem is this is not a bullet proof solution. It seems it may work, or it may not. The investigation is always difficult because it would require PESH password, few PMRs. From what I see, it can't be used as a resilient scenario. 

What is your experience, mainly IBMi shops? What is your VIOS resiliency strategy? I am not asking AIX community. For this group, the NIM server is a must. 

The idea is to reinstall the VIOS, then use rstviosbk. However, I didn't test this recovery plan as VIOS are very unlikely to fail completely ;)

 :) well this is not an excuse not to have the resiliency plan. I can give you two scenarios where we had to recover VIOS image.

a) new version of VIOS did not work properly. It was discovered immediately after VIO upgrade (SEA did not work - VIOS bug), and we had to recover from the backup

b) External storage failure - complete failure of full redundant enterprise external storage, with mirroring and all features - firmware bug. 

Restore of VIOS config, is not really option for us. We have dozen of VIOSes, on different POWER models, with different HW configuration. We do not have always the same VIOS release everywhere. It would be extremely difficult to restore VIOS image and then apply each specific fix on each for the particular VIOS. The RTO would be extremely prolonged (which is not an option in enterprise configuration).  

We have HW based HMCs everywhere, and dealing with PESH password is a must if you are dealing with misbehaving of  HMC software. 

https://www.ibm.com/support/pages/hmc-installios-cleanup

Hi,

Here, I have:

• scheduled "IO Configuration" for each VIOS, on 2 different HMC
• Regularly refreshed sysplans
• VIOS installation image at hand (customized for non-IBM storage)

The idea is to reinstall the VIOS, then use rstviosbk. However, I didn't test this recovery plan as VIOS are very unlikely to fail completely ;)

I used a very similar process to upgrade from PowerVM 3 to 4, as the automated upgrade won't work with our non-IBM storage. The only difference is that configuration was kept, I didn't need to restore it.

I'm not aware of PESH password nor PMRs, what is your hardware configuration?

Guys,

Does anyone use such functionality in production environment where it is regularly tested?

We want to get rid of NIMs and use "installios" HMC command which potentially should restore VIOS backup from a tar image. The problem is this is not a bullet proof solution. It seems it may work, or it may not. The investigation is always difficult because it would require PESH password, few PMRs. From what I see, it can't be used as a resilient scenario. 

What is your experience, mainly IBMi shops? What is your VIOS resiliency strategy? I am not asking AIX community. For this group, the NIM server is a must.