Rollback Procedures

When an upgrade encounters critical issues, a rollback to the previous version may be necessary. Energy Logserver does not provide an automated rollback mechanism — the process requires reinstalling the previous version and restoring configuration from a backup.

Prerequisites

Before every upgrade, the ./install.sh -u script offers to run configuration-backup.sh, which creates a backup archive in /root/. This backup contains configuration files, Elasticsearch indices data, templates, and cluster settings. Ensure this backup exists before attempting a rollback.

To verify that a backup archive exists:

ls -lt /root/backup-*.tar.gz | head -5

Procedure

1. Stop all services

systemctl stop logserver logserver-gui logstash alert cerebro \
  e-doc skimmer intelligence intelligence-scheduler license-service

2. Reinstall previous version

Use the installer package of the previous version:

tar xjf energylogserver-<previous-version>.x86_64.tar.bz2
cd install
./install.sh -u

3. Restore configuration from backup

Extract the backup archive created by configuration-backup.sh:

tar xzf /root/backup-<hostname>-<date>.tar.gz -C /tmp/restore

The archive contains the following configuration backups:

Archive file	Restore path
`elasticsearch_conf.tar.gz`	`/etc/elasticsearch/`
`kibana_conf.tar.gz`	`/etc/kibana/`
`logstash_conf.tar.gz`	`/etc/logstash/`
`alert_conf.tar.gz`	`/opt/alert/`
`cerebro_conf.tar.gz`	`/opt/cerebro/`
`skimmer_conf.tar.gz`	`/opt/skimmer/`
`intelligence_conf.tar.gz`	`/opt/intelligence/`
`license_conf.tar.gz`	`/opt/license-service/`
`e-doc_conf.tar.gz`	`/opt/e-doc/`
`logserver-store_conf.tar.gz`	`/etc/logserver-store/`

Restore only the configuration files that were affected by the upgrade. Each archive can be extracted individually, for example:

tar xzf /tmp/restore/<date>/elasticsearch_conf.tar.gz -C /
tar xzf /tmp/restore/<date>/kibana_conf.tar.gz -C /

4. Start services

Start the database service first and wait for it to become available:

systemctl start logserver
curl -u $USER:$PASSWORD "http://localhost:9200/_cluster/health?pretty"

Then start the remaining services:

systemctl start logserver-gui alert cerebro e-doc skimmer \
  intelligence intelligence-scheduler license-service logstash

5. Verify

systemctl status logserver logserver-gui logstash alert cerebro \
  e-doc skimmer intelligence intelligence-scheduler license-service \
  --no-pager