[Bug 1273462] Re: Users can mistakenly run init.d scripts and cause problems if an equivalent upstart job already exists
Steve Langasek
steve.langasek at canonical.com
Tue Sep 1 21:38:27 UTC 2015
I think the change description within the patch is wrong (or at least, misleading/incomplete). What this patch actually does is:
- for anything that calls /etc/init.d/foo, if /etc/init/foo.conf exists, make the call a passthrough to upstart instead of calling the init script.
- for anything that calls /etc/rc?.d/[KS]??foo, make it a no-op.
The *only* thing that's made a no-op is the invocation of the init
script *by the rc?.d symlink*. This is something that should never
happen anyway; our dependency-based boot is already supposed to bypass
invoking this script in favor of the upstart job. Any system that has
been configured to not use the dependency-based boot in trusty is going
to have race conditions at boot because of the combination of upstart
job and init script, and while making this change may change the
characteristics of the behavior at boot it would not be a regression
because things are already broken.
If anything interfaces with the /etc/init.d/foo script, that will still
function, except that now it will *work* by passing through to the
upstart job instead of trying to manipulate processes via (possibly non-
existent) pid files etc
--
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to upstart in Ubuntu.
https://bugs.launchpad.net/bugs/1273462
Title:
Users can mistakenly run init.d scripts and cause problems if an
equivalent upstart job already exists
Status in lsb package in Ubuntu:
Fix Released
Status in mysql-5.5 package in Ubuntu:
Invalid
Status in upstart package in Ubuntu:
Fix Released
Status in lsb source package in Trusty:
In Progress
Status in mysql-5.5 source package in Trusty:
Invalid
Status in upstart source package in Trusty:
Won't Fix
Status in lsb source package in Utopic:
Fix Released
Status in mysql-5.5 source package in Utopic:
Invalid
Status in upstart source package in Utopic:
Fix Released
Status in upstart package in Debian:
New
Bug description:
[ impact ]
Previously, init.d scripts that were replaced by upstart jobs had
"upstart-job" symlink as a redirect in-place, which directed users at
using upstart commands. Despite the good intentions, that never
actually taught people about the correct interfaces. Now with the
advent of co-installability of multiple init systems, users may have
systemd, upstart, and sysv-init all installed on users system and have
init.d scripts / upstart jobs / systemd units all available. To avoid
any doubt, we should support executing /etc/init.d/ scripts which may
call into upstart, or into systemd, or actually execute the script in
question depending on whether there is native configuration for that
particular job and which init system we are running under.
[ test case ]
Invoking init.d script should invoke upstart commands, for example:
$ /etc/init.d/ssh status
ssh start/running, process 4620
$ /etc/init.d/ssh stop
stop: Rejected send message, 1 matched rules; type="method_call", sender=":1.2469694" (uid=1000 pid=3908 comm="stop ssh ") interface="com.ubuntu.Upstart0_6.Job" member="Stop" error name="(unset)" requested_reply="0" destination="com.ubuntu.Upstart" (uid=0 pid=1 comm="/sbin/init")
$ sudo /etc/init.d/ssh stop
ssh stop/waiting
$ sudo /etc/init.d/ssh start
ssh start/running, process 5373
$ sudo /etc/init.d/ssh restart
ssh stop/waiting
ssh start/running, process 5405
Description: Ubuntu 13.10
Release: 13.10
mysql-server-5.5:
Installed: 5.5.35-0ubuntu0.13.10.1
Candidate: 5.5.35-0ubuntu0.13.10.1
Version table:
*** 5.5.35-0ubuntu0.13.10.1 0
500 http://us-east-1.ec2.archive.ubuntu.com/ubuntu/ saucy-updates/main amd64 Packages
500 http://security.ubuntu.com/ubuntu/ saucy-security/main amd64 Packages
100 /var/lib/dpkg/status
5.5.32-0ubuntu7 0
500 http://us-east-1.ec2.archive.ubuntu.com/ubuntu/ saucy/main amd64 Packages
In Ubuntu 13.10, the Upstart job and the init.d script do not work
properly. In previous versions, the init.d script was a symlink to
the wrapper script around upstart (/lib/init/upstart-job). This
conflict means that if the server was started using the init.d script,
upstart does not recognize that the server is running and will attempt
to start a second instance of mysqld.
Also problematic is that if the upstart job is started using the
service or start commands, the init.d script's "stop" function runs a
mysql shutdown, but upstart simply restarts mysqld (because it's
marked respawn in the upstart config).
Description: Ubuntu 14.04
Release: 14.04
mysql: mysql-server-5.5.43-0ubuntu0.14.04.1
The problem in some setup was that the upgrade von 12.04 to 14.04 requres the adjustment of the InnoDB log size. Therefore the start of MySQL via upstart failed directly while the one via init started successfully and then failed as below.
root at webserver01.kurt..ref:~# status mysql
mysql start/running, process 5866
root at webserver01.kurt..ref:~# /etc/init.d/mysql stop
* Stopping MySQL database server mysqld [ OK ]
root at webserver01.kurt..ref:~# status mysql
mysql start/running, process 6101
root at webserver01.kurt..ref:~# /etc/init.d/mysql status
* /usr/bin/mysqladmin Ver 8.42 Distrib 5.5.43, for debian-linux-gnu on x86_64
Copyright (c) 2000, 2015, Oracle and/or its affiliates. All rights reserved.
Oracle is a registered trademark of Oracle Corporation and/or its
affiliates. Other names may be trademarks of their respective
owners.
Server version 5.5.43-0ubuntu0.14.04.1-log
Protocol version 10
Connection Localhost via UNIX socket
UNIX socket /var/run/mysqld/mysqld.sock
Uptime: 7 sec
Threads: 1 Questions: 108 Slow queries: 0 Opens: 48 Flush tables: 1 Open tables: 41 Queries per second avg: 15.428
root at webserver01.kurt..ref:~# stop mysql
mysql stop/waiting
root at webserver01.kurt..ref:~# /etc/init.d/mysql status
* MySQL is stopped.
This is horrible. The "status" commands report the wrong status and the start/stop commands do not work. If our operators are not super careful, our orchestration and monitoring system will go wild, report the wrong status and/or perform continuous restarts of the system as they think the service is not running. So we also fix it in trusty. the result will looks as below:
ubuntu at maas:~/work/deb$ sudo start mysql
mysql start/running, process 8523
ubuntu at maas:~/work/deb$ sudo status mysql
mysql start/running, process 8523
ubuntu at maas:~/work/deb$ sudo /etc/init.d/mysql stop
mysql stop/waiting
ubuntu at maas:~/work/deb$ sudo status mysql
mysql stop/waiting
ubuntu at maas:~/work/deb$ sudo /etc/init.d/mysql status
mysql stop/waiting
[Regression Potential]
* Some scripts call '/etc/init.d/<service> reload' will not work if
upstart script lacks a reload function and the reload function of
/etc/init.d/<service> also uses PID. because it's managed by upstart
now (the PID file does not exist). We should enumerate all those bad
scripts and make them do the correct thing.
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/lsb/+bug/1273462/+subscriptions
More information about the foundations-bugs
mailing list