Issue uninstalling R81.20 jumbo 99

[the_rock]

Hey guys,

Wondering if someone might have an idea how to get around this problem. Client installed jumbo 99 on their mgmt and ever since then (week ago), every 2 days they have a problem where machine freezes totally and they cant open smart console or do much at all and it has to be rebooted. They tried removing jumbo, but it gives message I attached. We tried running installer uninstall and them fix name from clish, but shows fix not found, though it is there from cpinfo -yal command.

Any idea how to solve this?

Best,

Andy

[AkosBakos]

This sounds great before the Labour day 🙂 

First, a TAC case would be necessary here. 

Are the logs exported to SIEM? If yes -> consider about the revert to the earlier version. I know this is a rude step, but if thre is no log loss, only the rules must be recreated. (if there are no lot of changes)

I have only bad or worse tips except TAC.  😕

akos

[the_rock]

Köszönöm! Hope thats good translation lol

Anywho, yes, we opened TAC case, my colleague set one up yesterday, but still waiting for response. I know in old days of Check Point, there was ./UnixUninstall script, but thats been deprecated while ago...

Andy

[AkosBakos]

The translate is correct. Please keep us updated. About TAC: I had really good experience with Amber Paoli.

[the_rock]

Will let you know!

Andy

[gadt]

Hi the_rock,

please try 
autoupdatercli revert inext_nanoegg

Gadi

[the_rock]

That looks promising Gadi, we will try and will update you.

Andy

[the_rock]

Just tried it in my lab and no matter what I type, it keeps giving me below, though I see EXACT same fix in my R81.20 jumbo 88 mgmt lab.

Andy

[Expert@CP-MANAGEMENT:0]# autoupdatercli revert inext_nanoegg
Error: No product name for revert request, not reverting.
[Expert@CP-MANAGEMENT:0]# autoupdatercli revert BUNDLE_INEXT_NANO_EGG_AUTOUPDATE
Error: No product name for revert request, not reverting.
[Expert@CP-MANAGEMENT:0]# autoupdatercli revert INEXT_NANO_EGG_AUTOUPDATE Error: No product name for revert request, not reverting.
[Expert@CP-MANAGEMENT:0]# autoupdatercli revert INEXT_NANO_EGG
Error: No product name for revert request, not reverting.
[Expert@CP-MANAGEMENT:0]# autoupdatercli revert INEXT_NANOEGG
Error: No product name for revert request, not reverting.
[Expert@CP-MANAGEMENT:0]#

[gadt · Accepted solution]

Hi the_rock

last chance:
autoupdatercli revert inext_nanoegg_AutoUpdate

gadi

[the_rock]

Thank you, that worked!

Andy

[the_rock]

Hey @gadt 

So customer did exact same command, but though says its reverted, though log shows its not. This is why uninstall jumbo 99 is failing. Any idea if there is way to force it?

Andy

[gadt]

Hi @the_rock ,

I have approached relevant R&D for father assistance

Gadi

[the_rock]

Thanks a lot! Btw, we have remote with TAC today, so will update once done.

Andy

[Lesley]

Same issue on CLI? I see you tried the hotfix and not remove Jumbo from CLI is that right?

2. Connect to the command line on Gaia OS.

3. Log in to Gaia Clish.

4. Obtain the lock over the Gaia configuration database:

   HostName:0> lock database override

5. Uninstall the package:

   HostName:0> installer uninstall <Package_Number>

   The progress in percentage appears in Gaia Clish.

6. Server reboots automatically.

If posted steps do not help, there is indeed only TAC or rever via snapshot / backup. 

[the_rock]

Yep, as mentioned in my post, we tried that yesterday.

Andy

CP-MANAGEMENT> installer uninstall BUNDLE_INEXT_NANO_EGG_AUTOUPDATE
Info: Initiating uninstall of BUNDLE_INEXT_NANO_EGG_AUTOUPDATE...
Interactive mode is enabled. Press CTRL + C to exit (this will not stop the operation)
Result: Uninstall_Last_Take Failed
The package file is not in the repository.
CPUSE cannot find one of the uninstallation files. Contact Check Point Technical Services for further assistance.
CP-MANAGEMENT>

[gadt]

Hi the_rock

in case you did not see that message.
last chance:
autoupdatercli revert inext_nanoegg_AutoUpdate

Gadi

[the_rock · Accepted solution]

To add to all that @gadt has stated, here is something to make it even easier. Run autoupdatercli show and once you find right component name, then run autoupdatercli revert and then whatever component name is.

Example:

[Expert@CP-MANAGEMENT:0]# autoupdatercli revert core_uploader

Revert request handled. To see installation status, see logs: /opt/CPInstLog/AutoUpdater.log and /opt/CPInstLog/AutoUpdateLogs/core_uploader

No need to reboot and to be 100% sure it worked, you can run below (my example, but you gewt an idea)

autoupdatercli show | grep core_uploader

If nothing shows up, then you are good and know that fix was removed.

Best,

Andy

[gilne]

Hi @the_rock ,
please reach to me at gilne@checkpoint.com.
I am the owner of inext_nanoegg_AutoUpdate and i want to understand the exact scenario that you did in order to prevent it from happening in the future.

thanks.

[the_rock]

Hey @gilne 

I responded to your direct message, thank you! Btw, when we do autoupdatercli revert command, shows its handles, BUT, then you check the log and clearly states its not, very odd.

I cant figure out a way to force the removal of it.

Andy

[the_rock]

@gilne Sent you direct email on this.

My colleague and I truly believe jumbo did not install properly to begin with or was partial, as cpinfo -yall does NOT show it, but shows in web UI...we asked them to try install database from smart console and also try remove jumbo completely from web UI, rather than last jumbo, lets see if that works.