database revision control problem

[eldo37]

Hi all,

I have a problem to create database revision control with SPLAT on NGX R61.
When I reboot the management server I can create the database revision control but a few days later the database revision control installation failed.
I have the following error message :

Reactor::deregister_object: failed to delete object 0x0x9d5e3e8 from hash.
IVersionDBManager::CreateVersion with err 0x80004005 ('Unspecified error')
CCreateVersionHandler::CreateVersionMain: Failed to CreateVersion
Reactor::deregister_object: failed to delete object 0x0x99a3050 from hash.

I have check the mv_doc file. I have made a cpstop and cpstart.
I have delete all the contents of $FWDIR\conf\db_versions directory.

But I can only create the database revision control only one time after reboot then I received the error message.

Any idea ?

Thanks,
eldo.

[Porter]

how many revisions do you already have? we run into problems after we had lots of them, CP recommended to delete some of them, problem was gone afterwards

[eldo37]

Hi Porter,

I have 5 versions in $FWDIR\conf\db_versions\repository.

Before I had more versions but I rename db_versions with db_versions.ann


Thanks,
eldo

[Porter]

hm...we had about 200, looks like a different problem

[sdrapeau]

Hi eldo37,

Did you resolve your problem ?
Because I have the same with many installations

regards

[RayPesek]

Check out:
-------------------------------
Database Revision Control failed
Solution ID: #sk32010

Product: VPN-1 Pro (VPN-1/FW-1)
Version: NGX R61
Last Modified: 05-Dec-2006
Symptoms

* Error: "failed to create a version"
* fwm debug shows:
o CCreateVersionHandler::CreateVersion: Failed to IVersionDBManager::CreateVersion with err 0x80004005 ('Unspecified error')
o CCreateVersionHandler::CreateVersionMain: Failed to CreateVersion
------------------------------
It applies to SPLAT and Solaris, but it's a permissions issue so I'm not sure why it would work after a reboot. It may give you a hint, though.

Sometimes this error is caused by a missing InternalCA.C file in $FWDIR/conf or the $FWDIR\conf\mv_doc.C is zero bytes.

Ray

[sdrapeau]

All sk from CheckPoint was tested and verified with no success.

[dalfaro]

Hello.

I have the same problem and this was the solution:

- rename the db_versions folder
- close all GUI clients
- restart Smartcenter server (cprestart)
- login GUI Dashboard
- Create version

Enjoy

Regards from Chile.

[eldo37]

Hi Dalfaro,

Thank you for this solution.
it's work but if we reboot the gateway module we have the same problem.
" failed to create database revision control "

Any other idea ?

Thanks,
eldo

[melipla]

Ran into this problem the other day and have some info that may be beneficial to someone. First off, a little about what I saw--I create a revision when there's a policy push. That was failing so I tried going to File -> DB rev control -> Create revision, which worked. Then I could install policy and have it create a revision. Later in the day I changed between policies a few times and then wasn't able to create a revision any more, not even through the file create version path.

A call into Support, we ran through the multiples of SKs [most of which I had already done with no improvement], then started on the SK that specifies to enable debugging and then to do a database revision creation:

fw debug fwm on TDERROR_ALL_ALL=5
<attempt revision creation>
now the output is at $FWDIR/log/fwm.log look for a section like this:

[FWM 2625 2002595968]@hostname[18 May 19:03:48] Command is: create-mgmt-version
[FWM 2625 207600560]@hostname[18 May 19:03:48] do_links_getver: strncmp failed. Returning -2
[FWM 2625 207600560]@hostname[18 May 19:03:48] do_links_getver: strncmp failed. Returning -2
[FWM 2625 207600560]@hostname[18 May 19:03:48] do_links_getver: strncmp failed. Returning -2
[FWM 2625 207600560]@hostname[18 May 19:03:48] do_links_getver: strncmp failed. Returning -2
[FWM 2625 207600560]@hostname[18 May 19:03:48] do_links_getver: strncmp failed. Returning -2
[FWM 2625 207600560]@hostname[18 May 19:03:48] do_links_getver: strncmp failed. Returning -2
[FWM 2625 207600560]@hostname[18 May 19:03:48] do_links_getver: strncmp failed. Returning -2
[FWM 2625 207600560]@hostname[18 May 19:03:48] do_links_getver: strncmp failed. Returning -2
[FWM 2625 207600560]@hostname[18 May 19:03:48] do_links_getver: strncmp failed. Returning -2
CCreateVersionHandler::CreateVersion: Failed to IVersionDBManager::CreateVersion with err 0x80004005 ('Unspecified error')
CCreateVersionHandler::CreateVersionMain: Failed to CreateVersion

In my case, the "strncmp failed" relates to the firewall comparing data from the disk to data that is in memory. In order to correct this condition, the management server was rebooted and now I'm able to create new db versions. I had recently upgraded my management server to R65, it was up for less then a week when I started having this problem, I'm hopeful that it won't come back. The support person suggested that doing a cpstop/cpstart might correct the problem as well.

HTH

[lodown]

I ran in to a similar issue with R60A that was caused by either a bad version of gtar or gzip. Checkpoint sent me a replacement and it resolved the issue

-lodown

[RayPesek]

It seems to be just flat broked on R61. After taking the SmartConsole clients to R61 HFA01, the same as the SmartCenter (although the HFAs are unrelated), it doesn't occur as often now. But I still have to cpstop/cpstart the SmartCenter every week or so because it suddenly stops working. After a cpstop/cpstart, it works for another week or so.

Ray

[melipla]

My problem did come back. I reopened my SR and I received an updated version of gtar from Check Point (1.15.90) which, so far, has resolved my issue. I should note that there's more then one gtar located on my R65 SPLAT system. The one they had me update for db versions was in /opt/CPshrd-R65/util/.

[Seko07]

Hello,
I have been tested all solutions on this forum and all together on the web. I use NGX R62. But no has any return for this reason.
I research in continuously.
But, most solution is delete on /etc/fw/bin/conf db_version folder and than create new folder name is db_versions.
And then cprestart, this is a best solution.
So, no has any solution on the world.
If you are find any solutions on the Earth please contact with me on serkansaygun@gmail.com
Please.
Thnx.

[vigie]

I had this problem with a R65 SmartCenter.
I updated the gtar binary which is located in /opt/CPshrd-R65/util/ directory. This gtar binary was a 1.13.25 version, and I replaced it with a 1.15.1 one.
I found this gtar binary with rpmfind.net (and extract the rpm file).

And.. it works now !


Vincent G.

[Papinko]

Hai @ All,

just for information: We had the same problem with R65 on our SmartCenter after upgrade R55 => R65
We had already upgraded gtar but did it in /bin
Thanks Vincent for pointing out that gtar in /opt/CPshrd-R65/util/ has to be upgraded too!
BTW: We use gtar 1.18 and it works too.

Papinko

[ksims]

hello all,

I worked with support to use gtar in the /bin directory, they are aware this is an issue.

I renamed the $CPDIR/util/gtar file to a backup and created a link to the version of gtar located in /bin. Seems to be working just fine now.

thanks

[wgrixti]

I had the same issue. I had updated from NGX R60 to R65.

I created a symbolic link to the gtar file in /bin to replace the gtar file in $CPDIR/util.

Wayne

[melipla]

Beware of applying an HFA for R65 as it does not work with the non-default version of gtar. It is a "known issue"...

[Papinko]

Thank you for the hint, melipla. You should have posted that 2 days earlier ;-)
In the meantime we have restored our management station from backup, as we had changed first GTAR and then applied the HFA.

To go no risk, you should change to the original GTAR before appliing HFA (02 in our case) and switch to the new GTAR after the HFA is applied, right?

*wink* Papinko