VIFIB DESCENTRALIZED CLOUD COMPUTING

SlapOS is a decentralized Cloud Computing technology that can automate the deployment and configuration of applications in a heterogeneous environment.

Support of SlapOS prerm scripts

The goal of this document is to how the prerm script on SlapOS which runs when the partition is going to be destroyed

${table_of_content}

Introduction

Prerm extension on SlapOS was introduced from slapos node version 1.4.1. When the module is enabled on slapos server, this allow the partition to define a script which will be executed by slapgrid only when the partition is being destroyed.

How slapos prerm works

The software release should put the script to execute into the folder partition_root/etc/prerm. A good example of script is the one used in kvm which wipe vm images disk: https://lab.nexedi.com/nexedi/slapos.toolbox/blob/master/slapos/securedelete.py.

 The prerm script is invoked by slapgrid before destroy the partition. This is the sequence of partition destruction:

  1.  Slapgrid stop the partition: prepare the destruction, stop all processes first
  2. check retention delay lock: if there is a retention delay, which prevent partition destruction at the moment, stop here (no prerm script will run).
  3. Add all prerm scripts to supervisord and start them. They will be added in a new supervisord process group named 'slappartID-prerm' where ID is the number of the partition. To check all processes stated by prerm, use the command:

    slapos node status slappartID-prerm:

  4. check if prerm script are exited: when one of prerm scripts is running  the partition will not be destroyed, slapgrid will wait until they are all stopped or exited.
  5. destroy the partition: Slapgrid will cleanup supervisord and remove all files from the partition.

 

This is the sample of destruction of slappart98. The prerm script 'slapos_wipe_qemu_disk' is started here, then destruction process is stopped because we have a prerm script running. The script 'slapos_wipe_qemu_disk' continue to run in background.

[2017-10-05 10:55:11,900] INFO     Requested stop of slappart98...
[2017-10-05 10:55:11,950] INFO     Adding pre-delete scripts to supervisord...
[2017-10-05 10:55:11,952] DEBUG    Updating supervisord
[2017-10-05 10:55:12,018] INFO     Updated 'slappart98-prerm'
[2017-10-05 10:55:12,018] DEBUG    Supervisord updated
[2017-10-05 10:55:12,020] INFO     Starting pre-delete process u'slapos_wipe_qemu_disk'...
[2017-10-05 10:55:12,026] INFO     There are running processes into the partition, wait until they finish...

If we check partition's processes status, we have the result bellow. The wipe script is running.

slappart98:6tunnel-cluster-on-watch             STOPPED   Oct 05 02:10 PM
slappart98:bootstrap-monitor                    EXITED    Oct 05 01:55 PM
slappart98:certificate_authority-on-watch       STOPPED   Oct 05 02:10 PM
slappart98:crond-on-watch                       STOPPED   Oct 05 02:10 PM
slappart98:http-server-on-watch                 STOPPED   Oct 05 02:10 PM
slappart98:kvm-on-watch                         STOPPED   Oct 05 02:10 PM
slappart98:kvm_controller                       EXITED    Oct 05 01:55 PM
slappart98:monitor-httpd-graceful               EXITED    Oct 05 01:55 PM
slappart98:monitor-httpd-on-watch               STOPPED   Oct 05 02:10 PM
slappart98:websockify-on-watch                  STOPPED   Oct 05 02:10 PM
slappart98-prerm:slapos_wipe_qemu_disk          RUNNING   pid 3375, uptime 0:00:06

Then after the prerm script is finished, slapos node report will have this result:

[2017-10-05 11:25:07,891] INFO     Requested stop of slappart98...
[2017-10-05 11:25:07,957] INFO     Adding pre-delete scripts to supervisord...
[2017-10-05 11:25:07,959] DEBUG    Updating supervisord
[2017-10-05 11:25:08,003] DEBUG    Supervisord updated
[2017-10-05 11:25:08,007] INFO     Destroying Computer Partition slappart98...
[2017-10-05 11:25:08,015] DEBUG    Updating supervisord
[2017-10-05 11:25:08,048] INFO     Stopped 'slappart98'
[2017-10-05 11:25:08,049] INFO     Removed 'slappart98'
[2017-10-05 11:25:08,050] INFO     Stopped 'slappart98-prerm'
[2017-10-05 11:25:08,051] INFO     Removed 'slappart98-prerm'
[2017-10-05 11:25:08,051] DEBUG    Supervisord updated
[2017-10-05 11:25:08,453] INFO     Finished usage reports.

Now the partition is destroyed.

How to enable prerm extension

 Edit the slapos configuration in /etc/opt/slapos/slapos.cfg then add 'prerm' in manaer_list.

[slapos]

 computer_id = COMP-XXX

...

 manager_list = prerm

${related_subject_list}