dm-devel.redhat.com archive mirror
 help / color / mirror / Atom feed
From: "Benjamin Marzinski" <bmarzins@redhat.com>
To: John Stoffel <john@stoffel.org>
Cc: device-mapper development <dm-devel@redhat.com>,
	Christophe Varoqui <christophe.varoqui@gmail.com>
Subject: Re: [PATCH 01/17] multipathd: use /run instead of /var/run
Date: Tue, 29 Mar 2016 19:41:08 -0500	[thread overview]
Message-ID: <20160330004108.GG2915@octiron.msp.redhat.com> (raw)
In-Reply-To: <22266.35269.520687.195696@quad.stoffel.home>

On Tue, Mar 29, 2016 at 09:57:25AM -0400, John Stoffel wrote:
> 
> Benjamin> /var/run is now usually a symlink to /run.  If /var is on a separate
> Benjamin> filesytem, when multipathd starts up, it might end up writing to
> Benjamin> /var/run before the /var filesytem is mounted and thus not have its
> Benjamin> pidfile accessible at /var/run afterwards.  On most distrubutions /run
> Benjamin> is now a tmpfs and should always be available before multipathd is
> Benjamin> started, so multipath should just write there directly, instead of
> Benjamin> through the symlink.
> 
> I'm not sure I agree with this, it really smells more like a
> distribution problem, than a multipathd problem.  If multipathd starts
> up in an initramfs, how does it get reset to the proper /var/run or
> /run directory then?

Huh? /run is a tmpfs on almost all distros now. It gets remounted
when you switch the root from the initramfs, but it doesn't go away. So
there isn't a "proper" /run that appears later. /var/run just gets
symlinked to it. 

> And especially if it's a symlink, we shouldn't care at all either.

But /var/run isn't a symlink before /var is mounted.  That's the problem
that this is supposed to fix.

> What distros are using /run instead of /var/run these days?

Debian, Suse, Ubuntu, Fedora and some others. Pretty much everyone who
is using systemd made or is planning to make this switch. See
https://lists.fedoraproject.org/pipermail/devel/2011-March/150031.html

We could call it a distribution problem, but it will be a problem that
almost all distributions will face. To turn your question around, which
distribution do you think will be negatively impacted by this?

-Ben

> 
> John
> 
> 
> Benjamin> If /var/run is not a symlink, continue using it.
> 
> Benjamin> Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com>
> Benjamin> ---
> Benjamin>  Makefile.inc                    | 10 +++++++++-
> Benjamin>  libmultipath/defaults.h         |  2 +-
> Benjamin>  multipathd/multipathd.init.suse |  2 +-
> Benjamin>  3 files changed, 11 insertions(+), 3 deletions(-)
> 
> Benjamin> diff --git a/Makefile.inc b/Makefile.inc
> Benjamin> index c3ed73f..357dd33 100644
> Benjamin> --- a/Makefile.inc
> Benjamin> +++ b/Makefile.inc
> Benjamin> @@ -21,6 +21,14 @@ ifndef LIB
> Benjamin>  	endif
> Benjamin>  endif
>  
> Benjamin> +ifndef RUN
> Benjamin> +	ifeq ($(shell test -L /var/run -o ! -d /var/run && echo 1),1)
> Benjamin> +		RUN=run
> Benjamin> +	else
> Benjamin> +		RUN=var/run
> Benjamin> +	endif
> Benjamin> +endif
> Benjamin> +
> Benjamin>  ifndef SYSTEMD
> Benjamin>  	ifeq ($(shell systemctl --version > /dev/null 2>&1 && echo 1), 1)
> Benjamin>  		SYSTEMD = $(shell systemctl --version 2> /dev/null |  sed -n 's/systemd \([0-9]*\)/\1/p')
> Benjamin> @@ -54,7 +62,7 @@ ifndef RPM_OPT_FLAGS
> Benjamin>  endif
>  
> Benjamin>  OPTFLAGS     = $(RPM_OPT_FLAGS) -Wunused -Wstrict-prototypes
> Benjamin> -CFLAGS	     = $(OPTFLAGS) -fPIC -DLIB_STRING=\"${LIB}\"
> Benjamin> +CFLAGS	     = $(OPTFLAGS) -fPIC -DLIB_STRING=\"${LIB}\" -DRUN_DIR=\"${RUN}\"
> Benjamin>  SHARED_FLAGS = -shared
>  
> Benjamin>  %.o:	%.c
> Benjamin> diff --git a/libmultipath/defaults.h b/libmultipath/defaults.h
> Benjamin> index 8902f40..cf1d5be 100644
> Benjamin> --- a/libmultipath/defaults.h
> Benjamin> +++ b/libmultipath/defaults.h
> Benjamin> @@ -26,7 +26,7 @@
> Benjamin>  #define MAX_CHECKINT(a)		(a << 2)
>  
> Benjamin>  #define MAX_DEV_LOSS_TMO	0x7FFFFFFF
> Benjamin> -#define DEFAULT_PIDFILE		"/var/run/multipathd.pid"
> Benjamin> +#define DEFAULT_PIDFILE		"/" RUN_DIR "/multipathd.pid"
> Benjamin>  #define DEFAULT_SOCKET		"/org/kernel/linux/storage/multipathd"
> Benjamin>  #define DEFAULT_CONFIGFILE	"/etc/multipath.conf"
> Benjamin>  #define DEFAULT_BINDINGS_FILE	"/etc/multipath/bindings"
> Benjamin> diff --git a/multipathd/multipathd.init.suse b/multipathd/multipathd.init.suse
> Benjamin> index d1319b1..ed699fa 100644
> Benjamin> --- a/multipathd/multipathd.init.suse
> Benjamin> +++ b/multipathd/multipathd.init.suse
> Benjamin> @@ -17,7 +17,7 @@
>  
> Benjamin>  PATH=/bin:/usr/bin:/sbin:/usr/sbin
> Benjamin>  DAEMON=/sbin/multipathd
> Benjamin> -PIDFILE=/var/run/multipathd.pid
> Benjamin> +PIDFILE=/run/multipathd.pid
> Benjamin>  MPATH_INIT_TIMEOUT=10
> Benjamin>  ARGS=""
>  
> Benjamin> -- 
> Benjamin> 1.8.3.1
> 
> Benjamin> --
> Benjamin> dm-devel mailing list
> Benjamin> dm-devel@redhat.com
> Benjamin> https://www.redhat.com/mailman/listinfo/dm-devel

  reply	other threads:[~2016-03-30  0:41 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-03-29  3:12 [PATCH 00/17] Multipath patch sync Benjamin Marzinski
2016-03-29  3:12 ` [PATCH 01/17] multipathd: use /run instead of /var/run Benjamin Marzinski
2016-03-29 13:57   ` John Stoffel
2016-03-30  0:41     ` Benjamin Marzinski [this message]
2016-03-30 16:06       ` John Stoffel
2016-03-29  3:12 ` [PATCH 02/17] retrigger uevents to try and get the uid through udev Benjamin Marzinski
2016-03-29  3:13 ` [PATCH 03/17] Fix issues with user_friendly_names initramfs bindings Benjamin Marzinski
2016-03-29  3:13 ` [PATCH 04/17] Add libmpathcmd library and use it internally Benjamin Marzinski
2016-03-29  3:13 ` [PATCH 05/17] libmultipath: add ignore_new_boot_devs option Benjamin Marzinski
2016-03-29  3:13 ` [PATCH 06/17] libmultipath: fix PAD and PRINT macros Benjamin Marzinski
2016-03-29  3:13 ` [PATCH 07/17] libmultipath: Cut down on alua prioritizer ioctls Benjamin Marzinski
2016-03-29  3:13 ` [PATCH 08/17] multipathd: fail if pidfile can't be created Benjamin Marzinski
2016-03-29  3:13 ` [PATCH 09/17] libmultipath: check correct function for define Benjamin Marzinski
2016-03-29  3:13 ` [PATCH 10/17] multipathd: delay reloads during creation Benjamin Marzinski
2016-03-29 14:02   ` John Stoffel
2016-03-30  0:57     ` Benjamin Marzinski
2016-03-29  3:13 ` [PATCH 11/17] multipath: Fix minor text issues Benjamin Marzinski
2016-03-29  3:13 ` [PATCH 12/17] kpartx: verify partition devices Benjamin Marzinski
2016-03-29  3:13 ` [PATCH 13/17] multipath: add exclusive_pref_bit for alua prio Benjamin Marzinski
2016-03-29  3:13 ` [PATCH 14/17] multipathd: print "fail" when remove fails Benjamin Marzinski
2016-03-29  3:13 ` [PATCH 15/17] multipath: check partitions unused before removing Benjamin Marzinski
2016-03-29  3:13 ` [PATCH 16/17] multipathd.service: remove blk-availability Requires Benjamin Marzinski
2016-03-29  3:13 ` [PATCH 17/17] multipathd: use 64-bit int for command key Benjamin Marzinski
2016-04-07  2:10   ` multipathd: segfault in multipathd cli_add_map() Zhangguanghui

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20160330004108.GG2915@octiron.msp.redhat.com \
    --to=bmarzins@redhat.com \
    --cc=christophe.varoqui@gmail.com \
    --cc=dm-devel@redhat.com \
    --cc=john@stoffel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).