This MR brings some confort of use and of administration for the resiliency stack and resiliency related taks :

   *  logrotate some logs/feeds to save disk space
   * adds more promises on export and import instances
   * improves feedback on import failures :  promises are more verbose, more information on the takeover webpage, ...
   * better indempotency for the backup time information, which is now published instead of read from the .installed.cfg

Updated on 11/22 : 

   * improve pbs recipe : makes the trap command cross-shell
   * remove infinite loops in scripts generated by pbs, but run a given amount of time until raising an error (which can be catched by promises)
   * runner-importer script doesn't systematically fails if there is no requested SR.
   * adds promises for notifier feeds (on exporter and pbs) to check last result and check if feed is corrupted

This MR should be merged after slapos.core has been upgraded 

/reviewed-on nexedi/slapos!115

/reviewed-on nexedi/slapos!117

resilient: adds a promise to test integrity and the status of the last run of the feeds generated by notifier scripts

Instead, of that, the notifier wrappers will run the scripts several more times in case of failures
This is better than silent failures.

  Introduce NGINX on the same partition of apache to handle websocket\
  and eventsource types.

  The NGINX will run on another port and it would require a second ip at
  the machine for enable it.

  This configuration is a working version with fully https support, but
  some additional adjustments might be required.

Instead, of that, the notifier wrappers will run the scripts several more times in case of failures
This is better than silent failures.

  Use a single apache server to handle cache and normal apache configurations. With trafficserver,
  the access would follow the model:

    Apache > Traffic Server > Apache 'Cache' > Backend

  Now the configuration changed to:

    Apache > Traffic Server > Apache (same as before) > Backend

  This simplify the amount of apache process to manage on the frontend.

  Reduce a bit the number of sections also for create directories
  Move all 'set's and 'do's to earliest as possible, to keep buildout syntax more evident
  Drop duplicated logics and 'if's reducing general code

Otherwise, slaprunner-supervisord crashes because of unknown paths

Also, the instance-runner-export template becomes a Jinja template

Tests passing here:

https://nexedi.erp5.net/test_result_module/20161117-147BEEA9

/reviewed-on nexedi/slapos!118

because zope.interface is not required to build rebootstrap python.

This reverts commit ae808179.

Reason for revert: zodbtools depend on ZODB while currently neoppod &
erp5 are depending on `ZODB3 = 3.10` so upon installing both ZODB3 and
ZODB are installed. Together with our requirement

    stack/erp5/zope-versions.cfg:transaction = 1.1.1

This creates version conflict and installation fails:

2016-11-17 14:56:35,858 testsuite    : INFO     slapgrid_sr : 2016-11-17 14:56:35 slapos[3698] INFO The version, 1.1.1, is not consistent with the requirement, 'transaction>=1.6.1'.
2016-11-17 14:56:35,937 testsuite    : INFO     slapgrid_sr : 2016-11-17 14:56:35 slapos[3698] INFO While:
2016-11-17 14:56:35,937 testsuite    : INFO     slapgrid_sr : 2016-11-17 14:56:35 slapos[3698] INFO   Installing neoppod.
2016-11-17 14:56:35,937 testsuite    : INFO     slapgrid_sr : 2016-11-17 14:56:35 slapos[3698] INFO Error: Bad version 1.1.1

So revert zodbtools hooking for now until we upgrade our ZODB to ZODB > 3.

As advised by @Tyagov this adds zodbutils egg [1,2] to our software
stack: both NEO and ERP5 (since stack/erp5/ inherits from
neoppod/software-common.cfg)

I have not build-tested this patch but I hope it should be ok.

/cc @nexedi

[1] https://lab.nexedi.com/nexedi/zodbtools  
[2] https://pypi.python.org/pypi/zodbtools

/reviewed-on !116

This commit is a fix for 8787716d. It allows to get the random cron string
generated previously through parameters, and not reading manually .installed.cfg.

This message is also available via a promise

*  Includes a new recipe to get a random time
*  Moves generatepassword and generatemac into random.py
*  Fixes up some recipe, as the creation of a "random.py" file may collide
   with the import of the standard library random

This reverts commit 2b0c61d9.

If slapgrid fails, print the end of instance/software log in resilient log.
Now an the exit-code-file is always written, even if runner-importer fails.
The promise on this file gives the URL where the resilient log can
be accessed, to simplify the debugging