RE: Picking up development of dmraid

Mark-Willem Jansen <[email protected]> Thu, 19 Jul 2012 15:33:27 +0200
Newsgroups gmane.linux.ataraid
Message-ID <[email protected]>
--===============0718604269400541083==
Content-Type: multipart/alternative;
	boundary="_45f49b6b-b1b7-4a2a-a964-20d6dc3f0634_"

--_45f49b6b-b1b7-4a2a-a964-20d6dc3f0634_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable


Hi Heinz=2C

Concerning the partitioning support. I now feel that kpartx or (partx which=
 could be made compatible) would be the way to go in stead of dmraid. So it=
 would be a good idea to remove the partitioning support from dmraid.  This=
 will mean the package maintainer will need to make dmraid dependent on kpa=
rtx for it to work.

One remark on this: I found that on Debian to make kpartx work with dmraid =
during boot=2C one needs to make some changes to the multipath-tools packag=
es.

On a side note: Why does mdadm support MBR and GPT?

Concerning the usage of mdadm instead of dmraid. Me and probably and a larg=
e amount of AMD users will have a AMD chip-set which means a Promise FAKERA=
ID controller. So in my idea I had two options update dmraid to work with t=
he dm-raid target(probably the device-mapper target wrapper you are talking=
 about) to setup my RAID-5 or add support for the Promise metadata format t=
o mdadm. I picked the former as the code of dmraid look easier to understan=
d. =3B-)

If you say that adding a super-promise.c to mdadm is doable=2C I could chan=
ge my mind.

Just one last question I never really got an answer to. Can one use mdadm o=
n a dual boot system(MS and Linux) were the RAID partitions are shared? In =
other words will mdadm leave the metadata on the disks unchanged or in a st=
ate the the MS drivers can still recognize the RAID.

Would it be an idea to clean-up dmraid? Remove what is not needed anymore=
=2C or does not fit the purpose of the tool.

Kind regards=2C


Mark-Willem

Date: Thu=2C 19 Jul 2012 10:26:49 +0200
From: [email protected]
To: [email protected]
Subject: Re: Picking up development of dmraid


 =20
   =20
 =20
 =20
    On 07/18/2012 11:08 AM=2C Danny Wood wrote:

   =20
     =20
      Hi Mark-Willem Jansen

     =20

      You may want to speak with Phillip Susi of the Ubuntu Dmraid team.

      He built a set of patches a long while ago that I don't think got
      included in the stable dmraid.

      He knows the ins and outs of dmraid and has spends a lot of time
      bug fixing during Ubuntu release cycles.

     =20

      I think the main reason that this project has died is because it
      is a very niche market.

   =20
   =20

    Seconded WRT the niche market.

   =20

     It's
      usually only used by the people who run a dual boot with Windows
      as Mdadm is far superior for pure Linux installs.

   =20
   =20

    The later is exactly why things move to MD and eg. we're doing a
    device-mapper target wrapper

    to access the MD kernel runtime in order to make it accessible in
    LVM.

   =20

    Because the MD runtime has the long established field record it has=2C
    major FAKERAID OEMs decided

    to use it (namly Intel with their Intel Matrix RAID=2C isw in dmraid).

    mdadm gained external metadata format support along the lines of
    dmraid to allow for that and

    thus supports isw for long time now.

   =20

    As a result of that=2C Red Hat decided to not further develop dmraid.
    Actually we already asked publically=2C

    if dmraid is still mandatory to support the other metadata formats
    than DDF=2C Intel Matrix RAID and MD=2C

    which are supported by mdadm now anyway.

   =20

    No arguments it's still needed resulted from that so far.

   =20

    =20

      Also GPT can already be used on top of dmraid=2C as far as I know
      you use dmraid to initialise the block devs and kpartx to deal
      with partitions.

   =20
   =20

    There's no need to have code duplication for partitioinig support in
    another tool.

    For the record: the DOS partitioning support got added way back in
    time before kpartx addressed it

    (and never got pulled out).

   =20

    So use kpartx for activating partitionins on RAID sets.

   =20

   =20

    The most important question (as mentioned above) still persists
    though: is dmraid still needed or

    is any further development adequate  to support the Adaptec=2C
    Highpoint=2C Jmicron=2C LSI=2C NVidia=2C Promise=2C

    Silicon Image and VIA metadata formats? Are they still being used
    that much in the field or are users

    just happy with dmraid access to those in their mixed Linux/Windows
    environments?

    Requirement for pure Linux environment is MD/LVM anyway.

   =20

    We better get field feedback which we didn't get so far to answer
    that question.

   =20

    Heinz

   =20

    =20

      Good luck and best regards=2C

      Danny

     =20

     =20

      On 18/07/12 09:20=2C Mark-Willem Jansen
        wrote:

     =20
     =20
       =20
         Dear dmraid developers=2C

         =20

          Sometime in this mail-list it was said that the program dmraid
          was in maintaining mode and not further developed anymore. In
          the meantime the dm-developement team has put out new
          dm-target=2C which can be used by the tool.

         =20

          I would like to fork the latest RC and put on github=2C to
          continue developing the tool. I will give it a slightly new
          name=2C so people will not confuse it with the original. My plan
          is to add the support for new dm-targets and also implement
          more partition tables=2C starting with GPT.

         =20

          I am not really good at generating new names=2C but here are
          some ideas.

         =20

          dmraid-fbmw (forked by Mark-Willem)

          dmraid-fu (follow-up)

          dmraid-ext (extended version)

         =20

          So my question which name you think is a good one for the
          forked?

         =20

          And who can I connect if I have some questions about the tool.

         =20

          Greetings=2C

         =20

          Mark-Willem Jansen

         =20

       =20
       =20

       =20
       =20

        _______________________________________________
Ataraid-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/ataraid-list
     =20
     =20

     =20

     =20

     =20
     =20

      _______________________________________________
Ataraid-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/ataraid-list
   =20
   =20

 =20


_______________________________________________
Ataraid-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/ataraid-list 		 	   		  =

--_45f49b6b-b1b7-4a2a-a964-20d6dc3f0634_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

<html>
<head>
<style><!--
.hmmessage P
{
margin:0px=3B
padding:0px
}
body.hmmessage
{
font-size: 10pt=3B
font-family:Tahoma
}
--></style></head>
<body class=3D'hmmessage'><div dir=3D'ltr'>
Hi Heinz=2C<br><br>Concerning the partitioning support. I now feel that kpa=
rtx or (partx which could be made compatible) would be the way to go in ste=
ad of dmraid. So it would be a good idea to remove the partitioning support=
 from dmraid.&nbsp=3B This will mean the package maintainer will need to ma=
ke dmraid dependent on kpartx for it to work.<br><br>One remark on this: I =
found that on Debian to make kpartx work with dmraid during boot=2C one nee=
ds to make some changes to the multipath-tools packages.<br><br>On a side n=
ote: Why does mdadm support MBR and GPT?<br><br>Concerning the usage of mda=
dm instead of dmraid. Me and probably and a large amount of AMD users will =
have a AMD chip-set which means a Promise FAKERAID controller. So in my ide=
a I had two options update dmraid to work with the dm-raid target(probably =
the device-mapper target wrapper you are talking about) to setup my RAID-5 =
or add support for the Promise metadata format to mdadm. I picked the forme=
r as the code of dmraid look easier to understand. =3B-)<br><br>If you say =
that adding a super-promise.c to mdadm is doable=2C I could change my mind.=
<br><br>Just one last question I never really got an answer to. Can one use=
 mdadm on a dual boot system(MS and Linux) were the RAID partitions are sha=
red? In other words will mdadm leave the metadata on the disks unchanged or=
 in a state the the MS drivers can still recognize the RAID.<br><br>Would i=
t be an idea to clean-up dmraid? Remove what is not needed anymore=2C or do=
es not fit the purpose of the tool.<br><br>Kind regards=2C<br><br><br>Mark-=
Willem<br><br><div><div id=3D"SkyDrivePlaceholder"></div><hr id=3D"stopSpel=
ling">Date: Thu=2C 19 Jul 2012 10:26:49 +0200<br>From: [email protected]<br=
>To: [email protected]<br>Subject: Re: Picking up development of dmra=
id<br><br>
 =20
   =20
 =20
 =20
    On 07/18/2012 11:08 AM=2C Danny Wood wrote:<br>
    <blockquote cite=3D"mid:[email protected]">
     =20
      Hi Mark-Willem Jansen<br>
      <br>
      You may want to speak with Phillip Susi of the Ubuntu Dmraid team.<br=
>
      He built a set of patches a long while ago that I don't think got
      included in the stable dmraid.<br>
      He knows the ins and outs of dmraid and has spends a lot of time
      bug fixing during Ubuntu release cycles.<br>
      <br>
      I think the main reason that this project has died is because it
      is a very niche market.<br>
    </blockquote>
    <br>
    Seconded WRT the niche market.<br>
    <br>
    <blockquote cite=3D"mid:[email protected]"> It's
      usually only used by the people who run a dual boot with Windows
      as Mdadm is far superior for pure Linux installs.<br>
    </blockquote>
    <br>
    The later is exactly why things move to MD and eg. we're doing a
    device-mapper target wrapper<br>
    to access the MD kernel runtime in order to make it accessible in
    LVM.<br>
    <br>
    Because the MD runtime has the long established field record it has=2C
    major FAKERAID OEMs decided<br>
    to use it (namly Intel with their Intel Matrix RAID=2C isw in dmraid).<=
br>
    mdadm gained external metadata format support along the lines of
    dmraid to allow for that and<br>
    thus supports isw for long time now.<br>
    <br>
    As a result of that=2C Red Hat decided to not further develop dmraid.
    Actually we already asked publically=2C<br>
    if dmraid is still mandatory to support the other metadata formats
    than DDF=2C Intel Matrix RAID and MD=2C<br>
    which are supported by mdadm now anyway.<br>
    <br>
    No arguments it's still needed resulted from that so far.<br>
    <br>
    <blockquote cite=3D"mid:[email protected]"> <br>
      Also GPT can already be used on top of dmraid=2C as far as I know
      you use dmraid to initialise the block devs and kpartx to deal
      with partitions.<br>
    </blockquote>
    <br>
    There's no need to have code duplication for partitioinig support in
    another tool.<br>
    For the record: the DOS partitioning support got added way back in
    time before kpartx addressed it<br>
    (and never got pulled out).<br>
    <br>
    So use kpartx for activating partitionins on RAID sets.<br>
    <br>
    <br>
    The most important question (as mentioned above) still persists
    though: is dmraid still needed or<br>
    is any further development adequate&nbsp=3B to support the Adaptec=2C
    Highpoint=2C Jmicron=2C LSI=2C NVidia=2C Promise=2C<br>
    Silicon Image and VIA metadata formats? Are they still being used
    that much in the field or are users<br>
    just happy with dmraid access to those in their mixed Linux/Windows
    environments?<br>
    Requirement for pure Linux environment is MD/LVM anyway.<br>
    <br>
    We better get field feedback which we didn't get so far to answer
    that question.<br>
    <br>
    Heinz<br>
    <br>
    <blockquote cite=3D"mid:[email protected]"> <br>
      Good luck and best regards=2C<br>
      Danny<br>
      <br>
      <br>
      <div class=3D"ecxmoz-cite-prefix">On 18/07/12 09:20=2C Mark-Willem Ja=
nsen
        wrote:<br>
      </div>
      <blockquote cite=3D"mid:[email protected]">
        <style><!--
.ExternalClass .ecxhmmessage P
{padding:0px=3B}
.ExternalClass body.ecxhmmessage
{font-size:10pt=3Bfont-family:Tahoma=3B}

--></style>
        <div dir=3D"ltr"> Dear dmraid developers=2C<br>
          <br>
          Sometime in this mail-list it was said that the program dmraid
          was in maintaining mode and not further developed anymore. In
          the meantime the dm-developement team has put out new
          dm-target=2C which can be used by the tool.<br>
          <br>
          I would like to fork the latest RC and put on github=2C to
          continue developing the tool. I will give it a slightly new
          name=2C so people will not confuse it with the original. My plan
          is to add the support for new dm-targets and also implement
          more partition tables=2C starting with GPT.<br>
          <br>
          I am not really good at generating new names=2C but here are
          some ideas.<br>
          <br>
          dmraid-fbmw (forked by Mark-Willem)<br>
          dmraid-fu (follow-up)<br>
          dmraid-ext (extended version)<br>
          <br>
          So my question which name you think is a good one for the
          forked?<br>
          <br>
          And who can I connect if I have some questions about the tool.<br=
>
          <br>
          Greetings=2C<br>
          <br>
          Mark-Willem Jansen<br>
          <br>
        </div>
        <br>
        <fieldset class=3D"ecxmimeAttachmentHeader"></fieldset>
        <br>
        <pre>_______________________________________________
Ataraid-list mailing list
<a class=3D"ecxmoz-txt-link-abbreviated" href=3D"mailto:Ataraid-list@redhat=
.com">[email protected]</a>
<a class=3D"ecxmoz-txt-link-freetext" href=3D"https://www.redhat.com/mailma=
n/listinfo/ataraid-list" target=3D"_blank">https://www.redhat.com/mailman/l=
istinfo/ataraid-list</a></pre>
      </blockquote>
      <br>
      <br>
      <br>
      <fieldset class=3D"ecxmimeAttachmentHeader"></fieldset>
      <br>
      <pre>_______________________________________________
Ataraid-list mailing list
<a class=3D"ecxmoz-txt-link-abbreviated" href=3D"mailto:Ataraid-list@redhat=
.com">[email protected]</a>
<a class=3D"ecxmoz-txt-link-freetext" href=3D"https://www.redhat.com/mailma=
n/listinfo/ataraid-list" target=3D"_blank">https://www.redhat.com/mailman/l=
istinfo/ataraid-list</a></pre>
    </blockquote>
    <br>
 =20

<br>_______________________________________________
Ataraid-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/ataraid-list</div> 		 	   		  </div=
></body>
</html>=

--_45f49b6b-b1b7-4a2a-a964-20d6dc3f0634_--


--===============0718604269400541083==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
Ataraid-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/ataraid-list
--===============0718604269400541083==--