First Last Prev Next    No search results available      Search page      Enter new bug
Bug#: 177954
Alias:
Product:
Component:
Status: RESOLVED
Resolution: FIXED
Assigned To: Gentoo's Team for Core System packages <base-system@gentoo.org>
Hardware:
OS:
Version:
Priority:
Severity:
Reporter: Adam James <atj@pulsewidth.org.uk>
Add CC:
CC:
Remove selected CCs
URL:
Summary:
Status Whiteboard:
Keywords:

Filename Description Type Creator Created Size Actions
raid raid initscript for baselayout-2 text/plain Adam James 2007-05-10 18:15 0000 530 bytes Details
Create a New Attachment (proposed patch, testcase, etc.) View All

Bug 177954 depends on: Show dependency tree
Show dependency graph
Bug 177954 blocks:
Votes: 0    Show votes for this bug    Vote for this bug

Additional Comments: (this is where you put emerge --info)







View Bug Activity   |   Format For Printing   |   XML   |   Clone This Bug


Description:   Opened: 2007-05-10 18:12 0000
>=sys-apps/baselayout-2.0.0_alpha2 no longer directly initialises the volume addons located in /lib/rcscripts/addons.

The solution is to provide an initscript to be run at boot that starts the raid
addon.

Further background info in bugs #177100 and #175983.

------- Comment #1 From Adam James 2007-05-10 18:15:38 0000 -------
Created an attachment (id=118774) [edit]
raid initscript for baselayout-2

Had to add lvm to `before', else lvm would try to start before all the raid
arrays were initialised. Presumably this would be the same for other volume
addons like evms? 

------- Comment #2 From Petteri Räty 2007-05-16 20:24:50 0000 -------
(In reply to comment #1)
> 
> Had to add lvm to `before', else lvm would try to start before all the raid
> arrays were initialised. Presumably this would be the same for other volume
> addons like evms? 
> 

Or document RC_USE="raid" to conf.d/lvm 

------- Comment #3 From Roy Marples (RETIRED) 2007-05-16 20:32:09 0000 -------
Right, the volume init scripts should not depend on each other in the script.
At most provide defaults via RC_USE/RC_NEED/RC_BEFORE/RC_AFTER so the user can
decide the correct order.

------- Comment #4 From Adam James 2007-05-17 00:39:08 0000 -------
(In reply to comment #3)V
> Right, the volume init scripts should not depend on each other in the script.
> At most provide defaults via RC_USE/RC_NEED/RC_BEFORE/RC_AFTER so the user can
> decide the correct order.

How does VOLUME_ORDER from conf.d/volumes relate to this? It appears to define
an order in which the addons are started, yet I cannot see where this is
sourced during init.

------- Comment #5 From Roy Marples (RETIRED) 2007-05-17 06:35:15 0000 -------
(In reply to comment #4)
> How does VOLUME_ORDER from conf.d/volumes relate to this? It appears to define
> an order in which the addons are started, yet I cannot see where this is
> sourced during init.

It doesn't. volumes init and conf files no longer exist.

------- Comment #6 From Duncan 2007-08-15 00:21:18 0000 -------
Just upgraded to baselayout-2.0.0-rc2-r1 and ran into this myself.  Worked
around it temporarily with md= additions to the the kernel command line.  Now
that I'm up and running again, I was able to find this bug and simply copied
the lvm script to mdraid.

Suggestion:  The attachment here is called "raid".  Given the various kinds of
raid, that's ambiguous.  mdraid should be suitably more descriptive, IMO.

Question:  The attachment here is slightly different than the existing lvm
initscript (in rc2-r1).  Which style is current?

------- Comment #7 From Duncan 2007-08-24 11:31:26 0000 -------
(In reply to comment #6)
> Suggestion:  The attachment here is called "raid".  Given the various kinds
> of raid, that's ambiguous.  mdraid should be suitably more descriptive, IMO.

Of course, as I quickly discovered, as currently written with the ${SVCNAME}
placeholder, the addon script and the initscript must have the same name. <doh>
Still, mdraid for both seems appropriate.

------- Comment #8 From Robin Johnson 2007-10-03 09:09:08 0000 -------
mdadm-2.6.3-r1 in the tree with this fix.

First Last Prev Next    No search results available      Search page      Enter new bug