One year later…again millions of zero-byte files?

Almost a year ago I ran into the CPUSE timeout issue when Saving File Permissions. Last week I ran into a similar problem when creating a snapshot of the same VSX gateway. In a year time this VSX gateway received some newer Jumbo Hotfixes and at this point Jumbo Hotfix Take 286 (GA) was installed. I’ve never received a hotfix from TAC when I had the CPUSE issue back then.  We only had the workaround with manually deleting the temporary files in/opt/CPshrd-R77/CTX/CTX00001/tmp/. So might it be possible the snapshot and CPUSE issues are caused by the same problem?

In preparation of the installation of Jumbo Hotfix Take 292 I needed to create a snapshot first. Unfortunately it was taking forever and seemed to be stuck at 1%. I left the system alone for almost an hour and a half to see there was finally progress. At that point I remembered the CPUSE timeout issue and started searching for the temporary files. Again there were millions of zero-byte temporary files of which names were starting with ‘file’. Without a doubt this must be the same problem. Since I knew the workaround I’ve deleted them all to make sure I would not run into the CPUSE timeout issue.

I’ve added the find and delete commands as a comment to this topic on Check Point’s CheckMates community (by the way: this is an excellent community where you can find a lot of interesting stuff. I highly recommend it!).

But I could not believe the issue stil existed so I wondered if there would be a hotfix available. I found sk116679 titled ‘ Snapshot creation on Gaia OS is stuck at 1-2%’, released on 3 April 2017. It described the exact same zero-byte files. You need to be signed in to User Center (with the correct authorizations) to see there is a hotfix available. It’s not included in a Jumbo Hotfix (yet).

I’ve contacted TAC and to my surprise there was no hotfix available for the currently installed Jumbo Hotfix Take 286. I was going to install Jumbo Hotfix Take 292 (current GA) and asked if the hotfix was compatible. Again I received a rather disappointing answer. A portfix was needed. I asked the TAC engineer to remain the Service Request open as I was going to send a cpinfo as soon the new Jumbo Hotfix was installed. I want to have this issue fixed so we do not run into it again when installing another hotfixes or when creating snapshots.

My second request to the TAC engineer was to have this hotfix included in a future Jumbo Hotfix, otherwise we always have to request a portfix whenever a new Jumbo Hotfix is installed. The solution is probably not what this hotfix does. The mechanism that should clean up these temporary files needs to be fixed. The hotfix only addresses the symptoms and does nothing more than the workaround we found out a year ago.

Keep you updated as the story continues…