Bug 354055 – kernel - add support / help test raid1 new read_balance code

• Status: RESOLVED UPSTREAM
• Product: Gentoo Linux

[Description Roberto Spadim 2011-02-08 03:54:53 UTC]

hi guys, i made some changes to md raid1 software, could gentoo community test it? for me it work very nice =)
the raid1 new code is based in kernel 2.6.37
here is the new and old code:
www.spadim.com.br/raid1

just read_balance changed,4 modes: 
near_head(today) 
round_robin(counter per mirror) 
stripe (like raid0, with shift for make stripe less or more intensive)
time based (depending on head positioning time and read size it calculate the fasted mirror, some new improvement must be done but it works, improvements= get io queue of each mirror and make a time estimation of queue time, with it get more close best disk selection)

it´s open source please help me testing it, neil at raid kernel need some benchmarks to put it inside kernel source

for mixed speed mirrors time based is good
for ssd only round_robin is good (per mirror count, put 230 for 230mb/s, 150 for mb/s or multiples to make a good round robin)
for disks and/or ssd mixed or not, stripe is good (like raid0)
for disks only near head is good

[Comment 1 Roberto Spadim 2011-02-08 03:57:16 UTC]

Created attachment 261777 [details]
raid1.h  new source code

if your kernel is 2.6.37 just change raid1.c and raid1.h with raid1.new.c and raid1.new.h

[Comment 2 Roberto Spadim 2011-02-08 03:58:23 UTC]

Created attachment 261779 [details]
raid1.h  old source code

if your kernel is 2.6.37 just change raid1.c and raid1.h with raid1.new.c and raid1.new.h

[Comment 3 Roberto Spadim 2011-02-08 03:58:41 UTC]

Created attachment 261781 [details]
raid1.c  new source code

if your kernel is 2.6.37 just change raid1.c and raid1.h with raid1.new.c and raid1.new.h

[Comment 4 Roberto Spadim 2011-02-08 03:58:58 UTC]

Created attachment 261783 [details]
raid1.c  old source code

if your kernel is 2.6.37 just change raid1.c and raid1.h with raid1.new.c and raid1.new.h

[Comment 5 Petteri Räty (RETIRED) 2011-02-08 12:07:58 UTC]

The main purpose of this bug tracker is to handle issues within Gentoo that this does not seem to fall under. Also the Product chosen is entirely for non technical issues. To ask for interested testers you could try to see if there's an appropriate place to post on our forums:

http://forums.gentoo.org/

[Comment 6 Jorge Manuel B. S. Vicetto (RETIRED) 2011-02-09 02:37:55 UTC]

I'm opening the bug as this has nothing to do with DevRel.

As Petteri said this doesn't seem a valid gentoo bug, but if it were, it should probably be assigned to the kernel team.

[Comment 7 Jorge Manuel B. S. Vicetto (RETIRED) 2011-02-09 02:39:44 UTC]

Sorry, s/I'm opening/I'm dropping the restrictions on/

[Comment 8 Roberto Spadim 2011-02-09 03:28:52 UTC]

could place it as feature request? does gentoo have feature request list?

[Comment 9 Jorge Manuel B. S. Vicetto (RETIRED) 2011-02-09 03:40:28 UTC]

Let's see what the kernel team has to say about it.

[Comment 10 Jorge Manuel B. S. Vicetto (RETIRED) 2011-02-09 03:44:33 UTC]

@kernel:

Is there anything that you can do about this? Any suggestions to the developer on the best route to get his work tested and noticed by the Linux kernel team? If there's no way for this code to get to the gentoo ebuilds until this is handled upstream, can you please contact the user with details in case you choose to close the bug.

Roberto,
is that your goal? To be able to get your code into the Linux kernel?

[Comment 11 Mike Pagano 2011-03-17 10:52:22 UTC]

Please take this effort upstream. I suggest you post and look for assistance on the linux-raid mailing list.

http://vger.kernel.org/vger-lists.html#linux-raid

[Comment 12 Roberto Spadim 2011-03-17 13:10:59 UTC]

yeah i'm talking with neil
could anyone help me with a queue information? i want total number of writes/reads on queue buffer (ram memory queue), how could i get it? is it implemented or i should implement?