Create a Post
cancel
Showing results for 
Search instead for 
Did you mean: 
FXB
Contributor
Jump to solution

Config Migration of Gateway from R80.30 to R80.40?

Dear experts,

we are currently in the progress of planning the upgrade of our ClusterXL OpenServer Gateways from R80.30 to R80.40.

The SMS is already running on R80.40 for a few months.

 

We want to perform a fresh install of R80.40 on both of the OpenServer Gateways to take advantage of the new filesystem (so far we only did CPUSE upgrades in R80) and to make sure we have a "clean" state of the gateways.

In our test-lab we set up a R80.40 gateway with a fresh install and tried to import the config backup from one of the productive R80.30 gateways.

At first we got a warning about mismatching hotfixes and that the registry could not be updated, we could fix this by looking at the SK articles (which are really helpful).

However right now we get the message "Package is of version R80.30, which is incompatible with current version on machine R80.40", this is putting a hold on our upgrade plans.

Reading the documentation, it stats that backup restore might be support between versions, but is not 100% guaranteed, for what it seems for us, between R80.30 and R80.40 it is not possible.

Can you please give us suggestions on how we can backup and import the configuration on R80.40 after a fresh install?

Thank you very much.

Regards,

Franz

0 Kudos
1 Solution

Accepted Solutions
FXB
Contributor

Hi,

our solution for now is to fresh install a R80.30 system with an ext3 filesystem, import the backups of the current production gateways, upgrade to R80.40 Take 120 using CPUSE (now we are still at ext3 since no fresh install was made) and make and export a backup of the upgraded system with the production settings.

Than we go ahead and fresh install R80.40 Take 120 on the same test machine (with xfs now). Now we can import the backup with the production settings of the upgraded R80.40 system and make a backup. Now we have a backup with the production settings and xfs as filesystem which we will use for the production upgrade in a few weeks.

We repeat this procedure for all of the production gateways.

It is a bit of a overhead but this way we can assure to have backup that works on a fresh install R80.40 with the production settings for each gateways.

Maybe this helps for other people that experience the same problem.

Thanks for your help  @PhoneBoy .

Regards,

Franz

View solution in original post

0 Kudos
5 Replies
PhoneBoy
Admin
Admin

What version/JHF of R80.30?
And what kernel version?

If it’s the 3.10 variants of R80.30, I could see that not working properly.
That said, you should be able to export/import the Gaia config between the two.
This won’t get the manually configured files, of course.

FXB
Contributor

Hey,

thanks for your response.

We are currently running R80.30 Take 228 on both of the Gateways. Checking with "uname -a" we got the "3.10.0-693cpx86_64" kernel running, so yes it is indeed a 3.10 one.

From your response I am going to assume there are R80.30 Versions with a newer kernel - do you only get that by a fresh install? We upgraded from R80.10 up until R80.30 now via CPUSE. 

If we fresh install a R80.30 on our test-system, import our config, upgrade to R80.40 via CPUSE, export the config (so that we have a R80.40 config with productive settings), fresh install R80.40 on one of the productive node - the config import of the R80.40 should work while still having a fresh installed R80.40 i would assume. Can someone confirm this please?

Thank you very much.

Regards,

Franz

0 Kudos
PhoneBoy
Admin
Admin

The older kernel was 2.6.18, so you're already on the "newer" kernel.
And I guess you could upgrade some appliances to R80.30-3.10 with CPUSE, but not all.

If you fresh installed R80.40, or one of the 3.10 versions of R80.20/R80.30, the filesystem used is xfs.
In earlier releases, it's ext3, as noted here: https://sc1.checkpoint.com/documents/R80.40/WebAdminGuides/EN/CP_R80.40_RN/Topics-RN/Supported-Upgra... 
Since it requires different partitioning to utilize xfs, the only way to get xfs is a fresh install.
That different might explain why restoring a backup from R80.30 failed, but I'm not certain of that.

Of course, that might also mean when you attempt to restore the backup on a freshly installed R80.30 3.10 system (also uses the xfs filesystem), that will fail.
However, if that works, you should be able to upgrade it to R80.40.

0 Kudos
FXB
Contributor

Hi,

our solution for now is to fresh install a R80.30 system with an ext3 filesystem, import the backups of the current production gateways, upgrade to R80.40 Take 120 using CPUSE (now we are still at ext3 since no fresh install was made) and make and export a backup of the upgraded system with the production settings.

Than we go ahead and fresh install R80.40 Take 120 on the same test machine (with xfs now). Now we can import the backup with the production settings of the upgraded R80.40 system and make a backup. Now we have a backup with the production settings and xfs as filesystem which we will use for the production upgrade in a few weeks.

We repeat this procedure for all of the production gateways.

It is a bit of a overhead but this way we can assure to have backup that works on a fresh install R80.40 with the production settings for each gateways.

Maybe this helps for other people that experience the same problem.

Thanks for your help  @PhoneBoy .

Regards,

Franz

0 Kudos
PhoneBoy
Admin
Admin

There are clear benefits to being on xfs for management servers in terms of performance.
The benefits on a gateway are less clear, but might be helpful at some point in the future.

In any case, you usually don't need to do a full backup/restore of a gateway since the main OS configuration can be easily exported and most everything else comes from the management.

0 Kudos

Leaderboard

Epsum factorial non deposit quid pro quo hic escorol.

Upcoming Events

    CheckMates Events