Bug 1821228 - Automatic unlock of root filesystem doesn't happen, with tang server configured with FQDN
Summary: Automatic unlock of root filesystem doesn't happen, with tang server configur...
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Red Hat Gluster Storage
Classification: Red Hat Storage
Component: gluster-ansible
Version: rhgs-3.5
Hardware: x86_64
OS: Linux
unspecified
high
Target Milestone: ---
: RHGS 3.5.z Batch Update 2
Assignee: Gobinda Das
QA Contact: SATHEESARAN
URL:
Whiteboard:
Depends On:
Blocks: 1821216
TreeView+ depends on / blocked
 
Reported: 2020-04-06 10:56 UTC by SATHEESARAN
Modified: 2020-06-16 05:57 UTC (History)
6 users (show)

Fixed In Version: gluster-ansible-infra-1.0.4-8.el8rhgs,gluster-ansible-roles-1.0.5-8.el8rhgs
Doc Type: No Doc Update
Doc Text:
Clone Of: 1821216
Environment:
rhhiv, rhel8
Last Closed: 2020-06-16 05:57:32 UTC
Embargoed:


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Github gluster gluster-ansible-infra pull 92 0 None closed Support tang server configured with FQDN 2020-07-09 12:00:37 UTC
Red Hat Product Errata RHEA-2020:2575 0 None None None 2020-06-16 05:57:50 UTC

Description SATHEESARAN 2020-04-06 10:56:18 UTC
Description of problem:
------------------------
After successfully configuring Network Bound Disk Encryption(NBDE) using Clevis-Tang, automatic unlock of root filesystem doesn't happen, only in the case when FQDN was used for tang server

Version-Release number of selected component (if applicable):
-------------------------------------------------------------
gluster-ansible-infra-1.0.4-7

How reproducible:
-----------------
Always

Steps to Reproduce:
--------------------
1. Configure NBDE with RHHI-B, post running ansible playbook that configures the same, with tang server having a FQDN
2. Reboot the host

Actual results:
----------------
Host doesn't automatically unlocks, Fails with 'Error communicating with server'

Expected results:
-----------------
Automatically unlock root filesystem  and server should be up, without waiting for root encryption passphrase

--- Additional comment from SATHEESARAN on 2020-04-06 10:26:30 UTC ---

It was with the help of Sergio Correia, we could root cause this problem

This is primarily because in RHEL 8, both 'network' and 'network-legacy' modules are loaded in initramfs
<snip>
[root@localhost ~]# lsinitrd /boot/initramfs-4.18.0-192.el8.x86_64.img | grep -i network
network-legacy
network
kernel-network-modules
</snip>

when 'network-legacy' module is removed ( as in Fedora 31 ), and using 'network-manager' everything works fine

Thanks Sergio for your efforts in help in RCA

--- Additional comment from SATHEESARAN on 2020-04-06 10:30:46 UTC ---

To fix the problem, the dracut drop-in configuration file ( /etc/dracut.conf.d/clevis.conf)  needs to be updated with following lines:


omit_dracutmodules+="network-legacy"
add_dracutmodules+="clevis network-manager"

Along with the original content of 'clevis.conf' file the following is the content:

</snip>
omit_dracutmodules+="ifcfg"
omit_dracutmodules+="network-legacy"
add_dracutmodules+="clevis network-manager"
</snip>

Comment 3 SATHEESARAN 2020-04-17 17:14:29 UTC
Verified with gluster-ansible-infra-1.0.4-8.el8rhgs

1. Executed the playbook to setup NBDE, with DHCP based hostname for tang server
2. Post the playbook is executed successfully, reboot the host
3. Host could boot automatically without waiting passphrase prompt

Comment 5 errata-xmlrpc 2020-06-16 05:57:32 UTC
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHEA-2020:2575


Note You need to log in before you can comment on or make changes to this bug.