[PATCH 41/73] policycoreutils: mcstransd: write a pid file FIXME

[PATCH 41/73] policycoreutils: mcstransd: write a pid file FIXME

[Daniel J Walsh]

[-- Attachment #1: Type: text/plain, Size: 409 bytes --]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


   This patch looks good to me. acked.

systemd expects a pid file to be written by unit services.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk9x4YEACgkQrlYvE4MpobPLGwCgverG61aQYgrYV3P5BCISw+QO
ZP4An3ahmMsngAp0CxVDDbIirCGdLdzY
=wQig
-----END PGP SIGNATURE-----

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0041-policycoreutils-mcstransd-write-a-pid-file-FIXME.patch --]
[-- Type: text/x-patch; name="0041-policycoreutils-mcstransd-write-a-pid-file-FIXME.patch", Size: 0 bytes --]

Re: [PATCH 41/73] policycoreutils: mcstransd: write a pid file FIXME

[Colin Walters]

On Tue, 2012-03-27 at 11:49 -0400, Daniel J Walsh wrote:
> This patch looks good to me. acked.
> 
> systemd expects a pid file to be written by unit services.

It doesn't if the service is non-forking.  In fact moving to
non-forking services that don't write pid files is a major
improvement that systemd brings over sysvinit, because the
state is 100% reliable.  Pid files create the possibility
of desynchronization if the service crashes, but the pid
file remains.

So a better fix would be to change mcstransd to have a
--no-fork option, and update the unit file to use it.



--
This message was distributed to subscribers of the selinux mailing list.
If you no longer wish to subscribe, send mail to majordomo@tycho.nsa.gov with
the words "unsubscribe selinux" without quotes as the message.

Re: [PATCH 41/73] policycoreutils: mcstransd: write a pid file FIXME

[Daniel J Walsh]

[-- Attachment #1: Type: text/plain, Size: 1198 bytes --]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 03/27/2012 11:53 AM, Colin Walters wrote:
> On Tue, 2012-03-27 at 11:49 -0400, Daniel J Walsh wrote:
>> This patch looks good to me. acked.
>> 
>> systemd expects a pid file to be written by unit services.
> 
> It doesn't if the service is non-forking.  In fact moving to non-forking services that don't write pid files is a major improvement that systemd brings over sysvinit, because the state is 100% reliable.  Pid files create the possibility of desynchronization if the service crashes, but the pid file remains.
> 
> So a better fix would be to change mcstransd to have a --no-fork option, and update the unit file to use it.
> 
> 
> 
> -- This message was distributed to subscribers of the selinux mailing list. If you no longer wish to subscribe, send mail to majordomo@tycho.nsa.gov with the words "unsubscribe selinux" without quotes as the message.

How about. this patch.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk9yAMYACgkQrlYvE4MpobNgMwCgxJ0rYn5VJyybfV5OeuYNIrCU
CbUAnA5BoNnWx2/JXS5IzHgbfrn/IDJb
=FH0Q
-----END PGP SIGNATURE-----

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0143-Add-an-f-option-to-run-mcstransd-in-the-foreground.-.patch --]
[-- Type: text/x-patch; name="0143-Add-an-f-option-to-run-mcstransd-in-the-foreground.-.patch", Size: 0 bytes --]

Re: [PATCH 41/73] policycoreutils: mcstransd: write a pid file FIXME

[Colin Walters]

The naming here is confusing...the C variable should be called
"daemonize" or "do_fork" or something.

+       if (foreground && daemon(0, 0)) {

Would then be:

if (!foreground && daemon(0, 0))

Right?



--
This message was distributed to subscribers of the selinux mailing list.
If you no longer wish to subscribe, send mail to majordomo@tycho.nsa.gov with
the words "unsubscribe selinux" without quotes as the message.

Re: [PATCH 41/73] policycoreutils: mcstransd: write a pid file FIXME

[Daniel J Walsh]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 03/28/2012 12:24 PM, Colin Walters wrote:
> The naming here is confusing...the C variable should be called "daemonize" or "do_fork" or something.
> 
> +       if (foreground && daemon(0, 0)) {
> 
> Would then be:
> 
> if (!foreground && daemon(0, 0))
> 
> Right?
> 
> 
Ok using do_fork.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk9zSQIACgkQrlYvE4MpobNJeQCeMJJLClSc7KgtNxtEp9SRRgrm
gscAoJGil/oYdK+agV3OSDF/eUm8nixW
=0jm0
-----END PGP SIGNATURE-----

--
This message was distributed to subscribers of the selinux mailing list.
If you no longer wish to subscribe, send mail to majordomo@tycho.nsa.gov with
the words "unsubscribe selinux" without quotes as the message.