0

I have CentOS and Red hat enterprise Linux running, where I have now working Google Apps Engine and related other Python web applications.

Every week or depending on any Google Apps Engine latest release my working setup does not work any longer (right now its working but after one week it won't work again and its very weird problem without any answer from Google Apps Engine team members either).

For the time being the only solution for this is to rollback the whole CentOS/RHEL to 1 week back configuration including if kernel changes or anything related.

How can I tell CentOS/RHEL, go back to a installation point of one week back or to any restore point, so that it can reverse back to old setup when it was working for sure?

1
  • 2
    I have a lot of questions. Can you clarify exactly when the problem occurs? Is it every week? Exactly a week? (What day? Starting when?) Why do you think rolling back is a solution, let alone "the only solution"? And, where does Google Apps Engine fit into this? Do your apps running in the Google environment stop working? Or local test versions? Or do have a development environment (SDK, etc.) which stops working? What does "does not work any longer" specifically mean? What do you expect, and what happens instead? Are there any error messages?
    – mattdm
    Jul 23, 2013 at 2:22

1 Answer 1

2

You can try enabling yum's rollback feature as follows:

  1. vi /etc/yum.conf
  2. add this line to file: tsflags=repackage
  3. vi /etc/rpm/macros (create if non-existent)
  4. add this line to file: %_repackage_all_erasures 1

Now you can use rpm to rollback to different restore points:

$ rpm -Uvh –rollback ’21:00′
$ rpm -Uvh –rollback ’3 hours ago’
$ rpm -Uvh –rollback ‘august 13′
$ rpm -Uvh –rollback ‘yesterday’

All repackaged software is available here: /var/spool/repackage.

NOTE: You can only rollback from the point at which you enabled the above, you can't rollback prior to this!

References

You must log in to answer this question.