Issue with last upgrade tool package from R80.40 to R81.20

[patones1]

Hello

Last time I was testing an advanced upgrade from R80.40 version to R81.20, the last upgrade tools packet was downloaded and installed automatically (without clicking "Check for updates").
That should be OK if it worked correctly which it is not the case. When exporting the database, I got an error   "The current heap size of the CPM daemon is 1024 MB".

I followed the way to increase the size and I could export the database.

But it is not normal since I always was able to export the database with the old "Upgrade tools package" (ngm_upgrade_wrapper_997000619_1.tgz), without changing the heap size of the CPM daemon.

In my LAB I have no issue because I install the R80.40 without adding a gateway and I import manually the DA and the tools (ngm_upgrade_wrapper_997000619_1.tgz) from the CPUSE. But in production it is not the case since the management server, which is "on line",  has already installed the last upgrade tools.

What I need, for this case or future cases, is to be able to uninstall any "upgrade tools package". I need to have control over what I am going to use when proceeding to an advanced upgrade. Erasing the files under /var/log/CPda/repository  is not going to uninstall the upgrade tools already installed. I am not able to find by myself the way to uninstall any upgrade tool package.

Thanks for your help

Miguel

 

[Amir_Senn]

Hi,

AFAIK there's no automatic download of upgrade tools. Since we can migrate to GA or GA+JHF which don't have the same tools that makes sense in that manner.

I think that perhaps you've performed an operation that triggered installation of newer migration tool. For example, when you import upgrade package to CPUSE, you will have also have the relevant upgrade tools version with it installed. This is one case I'm aware of but there might be more cases.

Either way, you can check what your current tools using this: "cat /opt/CPupgrade-tools-R8*/version.info"

[Amir_Senn]

I stand corrected, tools are being updated.

[the_rock]

Personally, not aware of any way that can be done or t least supported way.

Andy

[AkosBakos]

Andy,

Maybe this helps 🙂

https://support.checkpoint.com/results/sk/sk169220

Akos

[the_rock]

Interesting...never seen that sk before.

Andy

[Tal_Paz-Fridman]

Hi Miguel,

Thanks for bringing this up! I understand your need, but each build includes important fixes, so we try to avoid users sticking with older versions.

For example, the heap size change was made to prevent Out of Memory issues during exports.

HTH

[patones1]

Thanks Tal_Paz,

I'll try an advanced  upgrade from R81.20 to R82 to see if it keeps asking for Heap size increase.
I hope that just clicking "check for updates" will get tools that work correctly (as usual). Like this I will avoid increasing heap size. If it is not the case, we will have to include the procedure (increasing heap size), as a routine one.

I`ll let you know the results.

On the other hand, it looks that there is no way to uninstall the upgrade tools already installed in case something goes wrong and I want to have more control over what tool package is installed.

Best regards

Miguel

[AkosBakos]

Hi,

Here is a link to sk, how to force a deployement agent install: 

https://support.checkpoint.com/results/sk/sk169220

You need to select the version what you vant to use, then install it. This will overwrite the the existing one.

Ákos

[patones1 · Accepted solution]

Hello,

Finally I successfully made an advanced upgrade from R81.20 to R82 without needing a Heap size increase.

I used the  upgrade tools downloaded when clicking "Check for Updates". Everything was fast and clean when exporting and importing the database.
So, there is no more issues.Thanks to everybody

Miguel

[AkosBakos]

Good to hear that, I had HEAP size issues on a VM last year. The quantity of heap size issues are growing day by day. This is my opinion.

If it is public, can you share the the VM resouces settings with us? RAM and CPU?

[patones1]

Below the resources of my VM. The Guest OS is "Other 3.x Linux (64 bits), which is normal because It was a R81.20 before upgrading.
What is weird is that in another VM, I installed directly the R82 without any issue using the same Guest OS (Other 3.x Linux (64 bits)) 

I used a disk of 110 GB where 20GB  were assigned for each partition (System-root and logs).

I am sharing some screenshots of the upgrade:

Before starting I clicked "Check for Updates", then I look at (/var/log/CPda/repository the upgrade tools are there

In $FWDIR/scripts/, after verifying, I launched the database export

Then I verified that migrate804811 is in /home/admin

Second part:
I imported the package Check_Point_R82_T777_Gaia_Install_and_Upgrade.tar and I did a clean install from CPUSE.

Once the install done, I clicked again in "Check for updates" to bring in the upgrade tools

migrate840811 file is still on /home/admin

Finally I imported the database

Connecting with the SmartConsole R82, I found all the rules and the SIC displayed an status of "Communicating"

As you see, it worked really well.

 

 

 

[the_rock]

Thanks for sharing that @patones1 ...super helpful! 👌

[the_rock]

Awesome work!