It has come to our attention that there are bugs in the CipherMail
update mechanism that prevent completing package update procedures.
While this issue has a low impact and can easily be worked around as
you can see below, a related issue results in database services failing
to restart on CipherMail cluster installations. This will cause node
downtime until the issue is manually resolved and can, if the regular
one-node-at-a-time update procedure is not followed, even lead to
downtime of the whole cluster.
The affected products are:
- CipherMail Gateway Professional/Enterprise 5.0.2 through 5.2.1.
- CipherMail Webmail 4.0.1 through 4.2.1.
The issues impact both the virtual appliances and CipherMail software
that was installed on existing Red Hat Enterprise Linux (RHEL) systems.
Installations of the CipherMail Community Gateway are not affected. The
issues have been present for some time but are triggered by recent
updates of the RHEL kernel and MariaDB packages.
Please use the following procedure to update your CipherMail
installations, or to repair a (partially) failed update. Always update
clusters one node at a time!
1. Connect to the CipherMail Console interface and choose
'File' -> 'Open shell' to get a shell.
2. Create an empty galera.cnf file (or overwrite the file installed by
the recent MariaDB update):
$ sudo sh -c 'echo "" > /etc/my.cnf.d/galera.cnf'
3. Cluster users: ensure that the database service is running:
$ sudo systemctl start mariadb
4. Update all packages if you haven't already:
$ sudo dnf update
5. Run or re-run the Ansible playbook:
$ sudo cm-run-playbook
6. Reboot the system:
$ sudo reboot
These issues will be fixed in the upcoming Gateway 5.2.2 and Webmail
4.2.2 releases. We would like to thank the customers that helped us
identify these issues and want to express our apologies for the
inconvenience that these bugs have caused.
If you have any questions or comments regarding these issues, please
feel free to reach out to our support team at support(a)ciphermail.com.
IT Automation Engineer