Bug 589775

Summary:	/sbin/start_udev should set a udev property for noiswmd kernel cmdline option
Product:	Red Hat Enterprise Linux 6	Reporter:	Hans de Goede <hdegoede>
Component:	udev	Assignee:	Harald Hoyer <harald>
Status:	CLOSED CURRENTRELEASE	QA Contact:	Karel Volný <kvolny>
Severity:	medium	Docs Contact:
Priority:	low
Version:	6.0	CC:	azelinka, harald, kay.sievers, kvolny, pknirsch
Target Milestone:	rc
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:	udev-147-2.18.el6	Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:	589774	Environment:
Last Closed:	2010-11-10 21:49:30 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Hans de Goede 2010-05-06 20:52:50 UTC

+++ This bug was initially created as a clone of Bug #589774 +++

For F-13 (and RHEL-6, I'll clone this bug), mdraid set activation has moved from rc.sysinit to udev rules. This includes the activation of isw / imsm BIOS RAID sets.

In order for the udev rules to be able to honor the noiswmd kernel cmdline option a udev property should be set when this option is present so that the udev rules can check for this.

The rules already check for rd_NO_MDIMSM, this was my idea / suggestion as
I assumed this property would be propagated from the initrd udev to the normal one, but it is not.

Still it would be the easiest to use rd_NO_MDIMSM as property name.

Comment 2 RHEL Program Management 2010-05-06 22:12:33 UTC

This request was evaluated by Red Hat Product Management for inclusion in a Red
Hat Enterprise Linux major release.  Product Management has requested further
review of this request by Red Hat Engineering, for potential inclusion in a Red
Hat Enterprise Linux Major release.  This request is not yet committed for
inclusion.

Comment 3 Kay Sievers 2010-05-30 22:04:42 UTC

We could make:
  IMPORT{cmdline}="rd_NO_MDIMSM"
work to be able to import specific variables from the kernel commandline during event processing.

This is a simple string only on the commandline, without a value, right? We would probably need to automatically translate that to:
  rd_NO_MDIMSM=1
because we test only for non-empty strings in udev so far, and handle empty values as unset variables.

Comment 4 Kay Sievers 2010-05-30 23:20:18 UTC

This might work:
  http://git.kernel.org/?p=linux/hotplug/udev.git;a=commitdiff;h=c4f6dcc4a5c774c4c5c60c7024d59081deecc7f8

Please check, or let me know if you have any better idea. Thanks!

Comment 5 Hans de Goede 2010-05-31 12:18:43 UTC

I like Kay's approach to this, unfortunately I don't have time atm to test this. But testing the basic functionality of IMPORT{cmdline} should be possible without the need for Intel BIOS RAID hardware access.

Comment 6 Kay Sievers 2010-05-31 12:22:26 UTC

(In reply to comment #5)
> But testing the basic functionality of IMPORT{cmdline} should be possible
> without the need for Intel BIOS RAID hardware access.    

Sure, I tested the basic things, and imported "quiet" and "root=", and this seems to work for me. :)

Comment 7 Hans de Goede 2010-05-31 13:20:57 UTC

(In reply to comment #6)
> (In reply to comment #5)
> > But testing the basic functionality of IMPORT{cmdline} should be possible
> > without the need for Intel BIOS RAID hardware access.    
> 
> Sure, I tested the basic things, and imported "quiet" and "root=", and this
> seems to work for me. :)    

Ok, then I see no reason why this would not work for no_iswmd.

Harald,

Could you do a new udev with this patch included (and also add it to Fedora) ? Then I'll do a patch to modify the mdadm udev rules, test it and send it to dledford

Thanks,

Hans

Comment 9 Karel Volný 2010-10-06 11:00:27 UTC

udev-147-2.29.el6

the file
/lib/udev/rules.d/65-md-incremental.rules
lists these rules:

ENV{rd_NO_MDIMSM}=="?*", GOTO="md_imsm_inc_end"
ENV{noiswmd}=="?*", GOTO="md_imsm_inc_end"

I've changed them to

ENV{rd_NO_MDIMSM}=="?*", PROGRAM="/usr/bin/logger 'bz589775: rd_NO_MDIMSM matched'", GOTO="md_imsm_inc_end"
ENV{noiswmd}=="?*", PROGRAM="/usr/bin/logger 'bz589775: noiswmd matched'", GOTO="md_imsm_inc_end"

now I get dozens of log messages:

# grep bz589775 /var/log/messages
Oct  6 12:47:08 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:08 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:47:16 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:08 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:08 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:08 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:08 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:08 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:08 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:08 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:08 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:08 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:08 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:08 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:09 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:10 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:10 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:20 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:20 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:20 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:20 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:20 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:20 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:20 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:20 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:20 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched
Oct  6 12:48:20 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched

after adding the parameter to kernel commandline:
# cat /proc/cmdline
ro root=/dev/mapper/vg_dhcp24235-lv_root rd_LVM_LV=vg_dhcp24235/lv_root rd_LVM_LV=vg_dhcp24235/lv_swap rd_NO_LUKS rd_NO_MD rd_NO_DM LANG=cs_CZ.UTF-8 SYSFONT=latarcyrheb-sun16 KEYBOARDTYPE=pc KEYTABLE=cz-lat2  rhgb quiet noiswmd

I still get just
Oct  6 12:54:19 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched

but *no* "bz589775: noiswmd matched"

=> so I'd say this does NOT work as expected ... could you confirm that my reproducer is correct, please?

Comment 10 Karel Volný 2010-10-06 11:04:04 UTC

um, call me "The Blind" - of course I can't get the second message if the first is matched and there is the GOTO ...

Comment 11 Harald Hoyer 2010-10-06 11:09:32 UTC

if "noiswmd" is found on the kernel command line, start_udev sets ENV{rd_NO_MDIMSM}==1 .

there is _no_  ENV{noiswmd}!!

Comment 12 Karel Volný 2010-10-06 11:20:35 UTC

ok, so I've swapped the order of the rules so "noiswmd" should be matched first -

ENV{noiswmd}=="?*", PROGRAM="/usr/bin/logger 'bz589775: noiswmd matched'", GOTO="md_imsm_inc_end"
ENV{rd_NO_MDIMSM}=="?*", PROGRAM="/usr/bin/logger 'bz589775: rd_NO_MDIMSM matched'", GOTO="md_imsm_inc_end"

but still I get only
Oct  6 13:08:46 dhcp-24-235 logger: bz589775: rd_NO_MDIMSM matched

... what have I missed this time?

Comment 13 Karel Volný 2010-10-06 11:23:49 UTC

(In reply to comment #11)
> if "noiswmd" is found on the kernel command line, start_udev sets
> ENV{rd_NO_MDIMSM}==1 .
> 
> there is _no_  ENV{noiswmd}!!

ah, that explains it

so the original rule

ENV{noiswmd}=="?*", GOTO="md_imsm_inc_end"

is wrong (can never be matched thus shouldn't be included)?

Comment 14 Harald Hoyer 2010-10-06 11:28:39 UTC

(In reply to comment #13)
> ENV{noiswmd}=="?*", GOTO="md_imsm_inc_end"
> 
> is wrong (can never be matched thus shouldn't be included)?

It is not wrong, it does not hurt, and it can be matched, if someone does:

/sbin/udevadm control --env=noiswmd=1

Comment 15 Karel Volný 2010-10-06 11:53:26 UTC

(In reply to comment #14)
> (In reply to comment #13)
> > ENV{noiswmd}=="?*", GOTO="md_imsm_inc_end"
> > 
> > is wrong (can never be matched thus shouldn't be included)?
> 
> It is not wrong, it does not hurt, and it can be matched, if someone does:
> 
> /sbin/udevadm control --env=noiswmd=1

seems like I'm missing the point here ... nevermind

so, appending "noiswmd" leads to setting "rd_NO_MDIMSM" which then matches the rule and skips the mdadm actions - which was the intent of this bug, if I get it right => VERIFIED

Comment 16 releng-rhel@redhat.com 2010-11-10 21:49:30 UTC

Red Hat Enterprise Linux 6.0 is now available and should resolve
the problem described in this bug report. This report is therefore being closed
with a resolution of CURRENTRELEASE. You may reopen this bug report if the
solution does not work for you.