[MDEV-10925] systemctl makes all versions of mariadn fail to start Created: 2016-09-29  Updated: 2016-09-30  Resolved: 2016-09-30

Status: Closed
Project: MariaDB Server
Component/s: Packaging
Affects Version/s: 10.2.2
Fix Version/s: N/A

Type: Bug Priority: Major
Reporter: Philip orleans Assignee: Unassigned
Resolution: Cannot Reproduce Votes: 0
Labels: None
Environment:

Centos 7 64



 Description   

Since a couple of days ago, in a Centos fully updated machine. all versions, 10.2 and 10.1, fail to start.
– Unit mariadb.service has begun starting up.
Sep 29 16:36:21 system55 systemd[795]: Failed at step NAMESPACE spawning /bin/sh: Invalid argument
– Subject: Process /bin/sh could not be executed
– Defined-By: systemd
This happens in both versions, 10.2 and 10.1, installed from scratch.
There is nothinhg wrong with mariadb, because this command works
mysqld --user=mysql



 Comments   
Comment by Elena Stepanova [ 2016-09-29 ]

What exactly happened "a couple of days ago" – did you upgrade MariaDB? Did you reconfigure something, for example your tmp dir?
Are you able to start other (non-MariaDB) services via systemd?

Comment by Philip orleans [ 2016-09-29 ]

I have Centos on automatic update. My guess is that Systemd got an update.
I installed version 5.5 and it starts, but I was working with version 10.X
I can give you access to the box and remove the current version 5.5.
The command systemctl start mriadb fails immediately.

Comment by Elena Stepanova [ 2016-09-29 ]

5.5 doesn't use systemd, it's irrelevant.
If you want to give us access to the machine, my public SSH key is here: https://launchpad.net/~elenst/+sshkeys

Comment by Philip orleans [ 2016-09-30 ]

No news about this? I had to start using Fedora 22 and go down one step in MariaDB version.
But somebody needs to fix this issue, it affects millions of users.

Comment by Elena Stepanova [ 2016-09-30 ]

philip_38,
Exactly the fact that we are not getting other similar complaints suggests that it's configuration-specific.
We can't have news unless you either provide the access to the machine as you earlier offered, or, for starters, answer the questions from my first comment.

Comment by Philip orleans [ 2016-09-30 ]

I don't want to post the IP here, kindly send me an email private. I already sent you one last night, but it was probably an old email of yours.
There are no more complaints because nobody upgraded Centos yet.
my email is venefax@gmail.com
Note: there is no configuration at all in my case. a blank new install, with all packets erased, reproduces the issue.

Yours
Federico

Comment by Elena Stepanova [ 2016-09-30 ]

Thanks, I've found your email now. I guess it didn't look legit enough so it went to spam. I'll check your instance shortly.

Comment by Philip orleans [ 2016-09-30 ]

Dear Elena
Kindly add the file /etc/yum.repos,d/maria.repo
for the latest 10.22 version, and install it
then simply
systemctl start mariadb

Comment by Elena Stepanova [ 2016-09-30 ]

philip_38,

I've just tried installing on your machine, I don't have a problem with either 10.1.18 or 10.2.2. I've left 10.2.2 installed, check it out.

Dependency Installed:
  MariaDB-client.x86_64 0:10.2.2-1.el7.centos                  MariaDB-common.x86_64 0:10.2.2-1.el7.centos                  MariaDB-compat.x86_64 0:10.2.2-1.el7.centos                 
 
Complete!
# systemctl start mariadb
# ps -ef | grep mysqld
mysql     7842     1  4 12:36 ?        00:00:00 /usr/sbin/mysqld
root      7874  6976  0 12:36 pts/11   00:00:00 grep --color=auto mysqld
# mysql -uroot -e "select @@version"
+----------------+
| @@version      |
+----------------+
| 10.2.2-MariaDB |
+----------------+
# 

Comment by Elena Stepanova [ 2016-09-30 ]

The reporter confirmed via email that the problem is gone; whatever the reason was, apparently it has already disappeared.

Generated at Thu Feb 08 07:45:58 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.