| Bug #21949 | RHEL4 installation bug | ||
|---|---|---|---|
| Submitted: | 31 Aug 2006 15:47 | Modified: | 12 Jun 2007 8:28 |
| Reporter: | Tom Ammon | Email Updates: | |
| Status: | No Feedback | Impact on me: | |
| Category: | MySQL Server: Installing | Severity: | S3 (Non-critical) |
| Version: | 5.0.24 | OS: | Linux (RedHat Enterprise Linux AS4) |
| Assigned to: | CPU Architecture: | Any | |
[31 Aug 2006 15:47]
Tom Ammon
[2 Sep 2006 10:11]
Valeriy Kravchuk
Thank you for a problem report. Please, specify the exact RPM used (or give URL to download). Version should be 5.0.x, not simply 5.0.
[3 Sep 2006 0:03]
Tom Ammon
The RPM I used is: MySQL-server-standard-5.0.24-0.rhel4.i386.rpm
[2 Oct 2006 17:25]
Joerg Bruehe
1) I checked the "spec" file:
In the "%post" section, it contains a line to add a user,
but intentionally ignores any errors:
useradd -M -r -d $mysql_datadir -s /bin/bash -c "MySQL server" -g %{mysqld_group} %{mysqld_user} 2> /dev/null || true
On uninstall, the user is intentionally *not* removed, so he should still exist from the previously installed 4.1.
2) I checked our RedHat 4 machine:
$ cat /etc/redhat-release
Red Hat Enterprise Linux AS release 4 (Nahant Update 2)
$ find /usr/bin /usr/sbin -name '*user*' -print | xargs ls -ld
...
lrwxrwxrwx 1 root root 7 Nov 24 2005 /usr/sbin/adduser -> useradd
...
-rwxr-x--- 1 root root 128312 Nov 17 2005 /usr/sbin/useradd
...
So I have no idea (yet) why it failed for the reporter.
Question:
Does that symlink between "useradd" and "adduser" exist, so that the spec file call to "useradd" and his suggested "adduser" lead to the same binary ?
[5 Oct 2006 9:53]
Valeriy Kravchuk
As you can see from Joerg's comment, we do have a command to add user. Can you try to repeat the same set of steps (install 4.1.x, upgrade to 5.0.x) but with checking of /etc/passwd, /etc/group and file prefissions (with numeric uid and gid) for the /var/lib/mysql directory?
[6 Nov 2006 0:00]
Bugs System
No feedback was provided for this bug for over a month, so it is being suspended automatically. If you are able to provide the information that was originally requested, please do so and change the status of the bug back to "Open".
[6 Nov 2006 10:58]
Valeriy Kravchuk
Sorry, I meant file permissions (not "file prefissions") in my previous question. It was a weird typo.
[7 Dec 2006 0:00]
Bugs System
No feedback was provided for this bug for over a month, so it is being suspended automatically. If you are able to provide the information that was originally requested, please do so and change the status of the bug back to "Open".
[7 Dec 2006 7:35]
Valeriy Kravchuk
Please, try to repeat with RPMs of a newer version, 5.0.27, and inform about the results.
[8 Jan 2007 0:00]
Bugs System
No feedback was provided for this bug for over a month, so it is being suspended automatically. If you are able to provide the information that was originally requested, please do so and change the status of the bug back to "Open".
[10 May 2007 19:18]
Tony Chamberlain
I had the same problem. The mysql rpms uninstall the earlier version and remove the mysql user. The new RPMs (listed below) do not add a mysql user back. The original mysql id number was 27. When I manually add user to add the mysql user back, it is 501. The stuff in /var/lib/mysql is still owned by 27. I notice that the original mysql init (in this case is actually mysqld) also does a chown -R mysqy:mysql on /var/lib/mysql when it starts. The new init (just mysql) does not. (Both are in /etc/init.d), so even when you add the mysql user back, the permissions aren't right until you do another chown. To summarize: 5.0.19 installation removes the mysql user and does not add it back (at least for me) manually adding the mysql user creates a different ID number (501 instead of 27) so /var/lib/mysql is still owned by 27 I made a thing to fix it. After running the mysql RPMs my script - checks for a mysql user. If one does not exist then it adds one. - adds a chown line to /etc/init.d/mysql under the 'start' portion - comments out the basedir= in /etc/my.cnf (suggested elsewhere). why is the previous mysql in /etc/init.d/mysqld and the new one in /etc/init.d/mysql ? Anyway here are the RPMs I used. I suspect the server is the important one. mysql++-2.2.3-1.el4.i386.rpm MySQL-client-standard-5.0.19-0.rhel4.i386.rpm MySQL-devel-standard-5.0.19-0.rhel4.i386.rpm MySQL-server-standard-5.0.19-0.rhel4.i386.rpm MySQL-shared-compat-5.0.19-0.rhel4.i386.rpm MySQL-shared-standard-5.0.19-0.rhel4.i386.rpm and centos 4.4 2.6.9-42.0.10.ELsmp #1 SMP
[12 May 2007 8:28]
Valeriy Kravchuk
Please, try to repeat with a newer version, 5.0.41, and inform about the results. Both 5.0.24 and 5.0.19 are really old.
[12 Jun 2007 23:00]
Bugs System
No feedback was provided for this bug for over a month, so it is being suspended automatically. If you are able to provide the information that was originally requested, please do so and change the status of the bug back to "Open".
