[Lustre-discuss] Kernel Patch Management
Andreas Dilger
adilger at sun.com
Sun Mar 9 23:36:12 PDT 2008
On Mar 10, 2008 11:21 +0530, ashok bharat bayana wrote:
> I'm trying to build lustre(1.6.4.2) version on 2.6.20.15 kernel on an x86 based machine.I'm using Ubuntu 7.10 flavour.
>
> While applying patches to the kernel in the series which_patch file,under the supported kernels, vanilla-2.6.20.15.series is not menioned but vanilla-2.6.18-.series is mentioned.
>
> So, can I proceed applying the series file of 2.6.18 version on a 2.6.20.15 version.
No, making Lustre kernel patches is not so easy. You should instead use
a supported kernel version for the server, and use patchless Lustre clients
for any systems which cannot run on a supported kernel.
> -----Original Message-----
> From: lustre-discuss-bounces at lists.lustre.org on behalf of lustre-discuss-request at lists.lustre.org
> Sent: Fri 2/29/2008 12:17 PM
> To: lustre-discuss at lists.lustre.org
> Subject: Lustre-discuss Digest, Vol 25, Issue 58
>
> Send Lustre-discuss mailing list submissions to
> lustre-discuss at lists.lustre.org
>
> To subscribe or unsubscribe via the World Wide Web, visit
> http://lists.lustre.org/mailman/listinfo/lustre-discuss
> or, via email, send a message with subject or body 'help' to
> lustre-discuss-request at lists.lustre.org
>
> You can reach the person managing the list at
> lustre-discuss-owner at lists.lustre.org
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Lustre-discuss digest..."
>
>
> Today's Topics:
>
> 1. Re: Configuration of lustre FS on single node
> (ashok bharat bayana)
> 2. Re: Configuration of lustre FS on single node (Amit Sharma)
> 3. Re: Configuration of lustre FS on single node
> (ashok bharat bayana)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Fri, 29 Feb 2008 12:07:18 +0530
> From: "ashok bharat bayana" <ashok.bharat.bayana at iiitb.ac.in>
> Subject: Re: [Lustre-discuss] Configuration of lustre FS on single
> node
> To: <lustre-discuss at lists.lustre.org>,
> <lustre-discuss at lists.lustre.org>
> Message-ID: <8626C1B7EB748940BCDD7596134632BE850229 at jal.iiitb.ac.in>
> Content-Type: text/plain; charset="iso-8859-1"
>
>
> Hello,
> I successfully build lustre(1.6.4.2) on my system for a patchless client But I dont know how to proceed in configuring the file system.
>
> I'm trying to build all MDS,client and MDT's on a single node.
> >From the tutorials I came to know that
>
> First, create an MDT for the "spfs" file system that uses the /dev/sda disk. This MDT also acts as the MGS for the site.
>
> $ mkfs.lustre --fsname spfs --mdt --mgs /dev/sda
>
> But by triggering this command my system crashed in which the whole data is lost and again I need to install the OS.
> I want help in proceeding of mounting a lustre file system.
>
> Thanks and Regards,
> Ashok Bharat
>
> -----Original Message-----
> From: lustre-discuss-bounces at lists.lustre.org on behalf of lustre-discuss-request at lists.lustre.org
> Sent: Wed 2/27/2008 10:30 PM
> To: lustre-discuss at lists.lustre.org
> Subject: Lustre-discuss Digest, Vol 25, Issue 56
>
> Send Lustre-discuss mailing list submissions to
> lustre-discuss at lists.lustre.org
>
> To subscribe or unsubscribe via the World Wide Web, visit
> http://lists.lustre.org/mailman/listinfo/lustre-discuss
> or, via email, send a message with subject or body 'help' to
> lustre-discuss-request at lists.lustre.org
>
> You can reach the person managing the list at
> lustre-discuss-owner at lists.lustre.org
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Lustre-discuss digest..."
>
>
> Today's Topics:
>
> 1. Multiple NICs per OST (Joshua Bower-Cooley)
> 2. obd error on MGS/MDT node (Jeremy Mann)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Tue, 26 Feb 2008 11:55:19 -0800
> From: Joshua Bower-Cooley <jbowercooley at lcogt.net>
> Subject: [Lustre-discuss] Multiple NICs per OST
> To: lustre-discuss at lists.lustre.org
> Message-ID: <200802261155.19907.jbowercooley at lcogt.net>
> Content-Type: text/plain; charset="utf-8"
>
> Hi-
> I've had lustre running happily for some time over a single 10g ethernet NIC
> per node. After switching to dual 10g and creating a new filesystem
> (1.6.4.2), I'm seeing nothing but keep-alive packets with bad checksums.
>
> What is the current "correct" way to do this now? The manual suggest not using
> bonding, but several list postings now reccommend it. Without bonding, do I
> need to have my 2 switches stacked, or will Lustre recognize the division in
> my subnet?
>
> LNET module options I've tried are:
> 1) networks="tcp0(eth2,eth3)"
> 2) ip2nets="tcp(eth2,eth3); tcp(eth2) 10.9.[1-4].*; tcp(eth3) 10.9.[5-8].*;"
> and many other variations
>
> All other services are functional on both interfaces.
>
> Thanks in advance,
> J Bower-Cooley
> Scientific Systems Engineer
> Las Cumbres Observatory Global Telescope
>
>
> ------------------------------
>
> Message: 2
> Date: Wed, 27 Feb 2008 09:50:08 -0600 (CST)
> From: "Jeremy Mann" <jeremy at biochem.uthscsa.edu>
> Subject: [Lustre-discuss] obd error on MGS/MDT node
> To: lustre-discuss at lists.lustre.org
> Message-ID:
> <62901.24.173.62.147.1204127408.squirrel at biochem.uthscsa.edu>
> Content-Type: text/plain;charset=iso-8859-1
>
> Today we are starting to see obd_change_cbdata errors on our MGS/MDT node.
> What does this error mean?
>
> Version of lustre and kernel is:
>
> 2.6.9-42.0.10.EL_lustre-1.6.0.1smp X86_64
>
> LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata()) Skipped 1
> previous similar message
> LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata())
> obd_change_cbdata: NULL export
> LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata()) Skipped 1
> previous similar message
> LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata())
> obd_change_cbdata: NULL export
> LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata()) Skipped 2
> previous similar messages
> LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata())
> obd_change_cbdata: NULL export
> LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata()) Skipped 4
> previous similar messages
> LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata())
> obd_change_cbdata: NULL export
> LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata()) Skipped 5
> previous similar messages
> LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata())
> obd_change_cbdata: NULL export
> LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata()) Skipped 1
> previous similar message
>
> --
> Jeremy Mann
> jeremy at biochem.uthscsa.edu
>
> University of Texas Health Science Center
> Bioinformatics Core Facility
> http://www.bioinformatics.uthscsa.edu
> Phone: (210) 567-2672
>
>
>
> ------------------------------
>
> _______________________________________________
> Lustre-discuss mailing list
> Lustre-discuss at lists.lustre.org
> http://lists.lustre.org/mailman/listinfo/lustre-discuss
>
>
> End of Lustre-discuss Digest, Vol 25, Issue 56
> **********************************************
>
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: http://lists.lustre.org/pipermail/lustre-discuss/attachments/20080229/887613dd/attachment-0001.html
>
> ------------------------------
>
> Message: 2
> Date: Fri, 29 Feb 2008 12:03:09 +0530
> From: Amit Sharma <Amit.Sharma at Sun.COM>
> Subject: Re: [Lustre-discuss] Configuration of lustre FS on single
> node
> To: ashok bharat bayana <ashok.bharat.bayana at iiitb.ac.in>
> Cc: lustre-discuss at lists.lustre.org
> Message-ID: <47C7A725.5050309 at sun.com>
> Content-Type: text/plain; format=flowed; charset=ISO-8859-1
>
> Ashok,
>
> Could you let us know what are the modules that you have installed on
> the system. The linux version on your system. And the steps that you
> have done so far.
>
> thanks,
> Amit
> ashok bharat bayana wrote:
> >
> > Hello,
> > I successfully build lustre(1.6.4.2) on my system for a patchless client
> > But I dont know how to proceed in configuring the file system.
> >
> > I'm trying to build all MDS,client and MDT's on a single node.
> > >From the tutorials I came to know that
> >
> > First, create an MDT for the "spfs" file system that uses the /dev/sda
> > disk. This MDT also acts as the MGS for the site.
> >
> > $ mkfs.lustre --fsname spfs --mdt --mgs /dev/sda
> >
> > But by triggering this command my system crashed in which the whole data
> > is lost and again I need to install the OS.
> > I want help in proceeding of mounting a lustre file system.
> >
> > Thanks and Regards,
> > Ashok Bharat
> >
> > -----Original Message-----
> > From: lustre-discuss-bounces at lists.lustre.org on behalf of
> > lustre-discuss-request at lists.lustre.org
> > Sent: Wed 2/27/2008 10:30 PM
> > To: lustre-discuss at lists.lustre.org
> > Subject: Lustre-discuss Digest, Vol 25, Issue 56
> >
> > Send Lustre-discuss mailing list submissions to
> > lustre-discuss at lists.lustre.org
> >
> > To subscribe or unsubscribe via the World Wide Web, visit
> > http://lists.lustre.org/mailman/listinfo/lustre-discuss
> > or, via email, send a message with subject or body 'help' to
> > lustre-discuss-request at lists.lustre.org
> >
> > You can reach the person managing the list at
> > lustre-discuss-owner at lists.lustre.org
> >
> > When replying, please edit your Subject line so it is more specific
> > than "Re: Contents of Lustre-discuss digest..."
> >
> >
> > Today's Topics:
> >
> > 1. Multiple NICs per OST (Joshua Bower-Cooley)
> > 2. obd error on MGS/MDT node (Jeremy Mann)
> >
> >
> > ----------------------------------------------------------------------
> >
> > Message: 1
> > Date: Tue, 26 Feb 2008 11:55:19 -0800
> > From: Joshua Bower-Cooley <jbowercooley at lcogt.net>
> > Subject: [Lustre-discuss] Multiple NICs per OST
> > To: lustre-discuss at lists.lustre.org
> > Message-ID: <200802261155.19907.jbowercooley at lcogt.net>
> > Content-Type: text/plain; charset="utf-8"
> >
> > Hi-
> > I've had lustre running happily for some time over a single 10g ethernet NIC
> > per node. After switching to dual 10g and creating a new filesystem
> > (1.6.4.2), I'm seeing nothing but keep-alive packets with bad checksums.
> >
> > What is the current "correct" way to do this now? The manual suggest not
> > using
> > bonding, but several list postings now reccommend it. Without bonding, do I
> > need to have my 2 switches stacked, or will Lustre recognize the division in
> > my subnet?
> >
> > LNET module options I've tried are:
> > 1) networks="tcp0(eth2,eth3)"
> > 2) ip2nets="tcp(eth2,eth3); tcp(eth2) 10.9.[1-4].*; tcp(eth3) 10.9.[5-8].*;"
> > and many other variations
> >
> > All other services are functional on both interfaces.
> >
> > Thanks in advance,
> > J Bower-Cooley
> > Scientific Systems Engineer
> > Las Cumbres Observatory Global Telescope
> >
> >
> > ------------------------------
> >
> > Message: 2
> > Date: Wed, 27 Feb 2008 09:50:08 -0600 (CST)
> > From: "Jeremy Mann" <jeremy at biochem.uthscsa.edu>
> > Subject: [Lustre-discuss] obd error on MGS/MDT node
> > To: lustre-discuss at lists.lustre.org
> > Message-ID:
> > <62901.24.173.62.147.1204127408.squirrel at biochem.uthscsa.edu>
> > Content-Type: text/plain;charset=iso-8859-1
> >
> > Today we are starting to see obd_change_cbdata errors on our MGS/MDT node.
> > What does this error mean?
> >
> > Version of lustre and kernel is:
> >
> > 2.6.9-42.0.10.EL_lustre-1.6.0.1smp X86_64
> >
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata()) Skipped 1
> > previous similar message
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata())
> > obd_change_cbdata: NULL export
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata()) Skipped 1
> > previous similar message
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata())
> > obd_change_cbdata: NULL export
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata()) Skipped 2
> > previous similar messages
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata())
> > obd_change_cbdata: NULL export
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata()) Skipped 4
> > previous similar messages
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata())
> > obd_change_cbdata: NULL export
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata()) Skipped 5
> > previous similar messages
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata())
> > obd_change_cbdata: NULL export
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata()) Skipped 1
> > previous similar message
> >
> > --
> > Jeremy Mann
> > jeremy at biochem.uthscsa.edu
> >
> > University of Texas Health Science Center
> > Bioinformatics Core Facility
> > http://www.bioinformatics.uthscsa.edu
> > Phone: (210) 567-2672
> >
> >
> >
> > ------------------------------
> >
> > _______________________________________________
> > Lustre-discuss mailing list
> > Lustre-discuss at lists.lustre.org
> > http://lists.lustre.org/mailman/listinfo/lustre-discuss
> >
> >
> > End of Lustre-discuss Digest, Vol 25, Issue 56
> > **********************************************
> >
> >
> > ------------------------------------------------------------------------
> >
> > _______________________________________________
> > Lustre-discuss mailing list
> > Lustre-discuss at lists.lustre.org
> > http://lists.lustre.org/mailman/listinfo/lustre-discuss
>
>
> --
> Amit Sharma
> Lustre Engineering
> Sun Microsystems, Bangalore.
>
>
> ------------------------------
>
> Message: 3
> Date: Fri, 29 Feb 2008 12:27:53 +0530
> From: "ashok bharat bayana" <ashok.bharat.bayana at iiitb.ac.in>
> Subject: Re: [Lustre-discuss] Configuration of lustre FS on single
> node
> To: "Amit Sharma" <Amit.Sharma at Sun.COM>
> Cc: lustre-discuss at lists.lustre.org
> Message-ID: <8626C1B7EB748940BCDD7596134632BE85022A at jal.iiitb.ac.in>
> Content-Type: text/plain; charset="iso-8859-1"
>
>
> Hi,
> the kernel version I used is 2.6.20.15 and the distro is Ubuntu 7.10
> The steps so far I have done are:
> 1)I have downloaded the sources lustre(1.6.4.2) from CFS
> 2)I tried to build lustre for a patchless client in which i triggered the command,
> [lustre]$ ./configure --with-linux=/unpatched/kernel/source
> 3)then I build it using make and then installed the modules using make install commands
> (I'm not sure whether the modules are installed correctly or not.Isn't the modules are installed using make install cmd)
> 4)Now in /usr/lib there is a lustre folder created.
> (Can we setup all MDS,client and OST's on a single system? my system configuration is Intel Pentium 512 MB RAM 1.6 GHz)
> Could you please tell me how to mount lustre fs in my laptop.
>
> Thanks and Regards,
> Ashok Bharat
> -----Original Message-----
> From: Amit Sharma [mailto:Amit.Sharma at Sun.COM]
> Sent: Fri 2/29/2008 12:03 PM
> To: ashok bharat bayana
> Cc: lustre-discuss at lists.lustre.org
> Subject: Re: [Lustre-discuss] Configuration of lustre FS on single node
>
> Ashok,
>
> Could you let us know what are the modules that you have installed on
> the system. The linux version on your system. And the steps that you
> have done so far.
>
> thanks,
> Amit
> ashok bharat bayana wrote:
> >
> > Hello,
> > I successfully build lustre(1.6.4.2) on my system for a patchless client
> > But I dont know how to proceed in configuring the file system.
> >
> > I'm trying to build all MDS,client and MDT's on a single node.
> > >From the tutorials I came to know that
> >
> > First, create an MDT for the "spfs" file system that uses the /dev/sda
> > disk. This MDT also acts as the MGS for the site.
> >
> > $ mkfs.lustre --fsname spfs --mdt --mgs /dev/sda
> >
> > But by triggering this command my system crashed in which the whole data
> > is lost and again I need to install the OS.
> > I want help in proceeding of mounting a lustre file system.
> >
> > Thanks and Regards,
> > Ashok Bharat
> >
> > -----Original Message-----
> > From: lustre-discuss-bounces at lists.lustre.org on behalf of
> > lustre-discuss-request at lists.lustre.org
> > Sent: Wed 2/27/2008 10:30 PM
> > To: lustre-discuss at lists.lustre.org
> > Subject: Lustre-discuss Digest, Vol 25, Issue 56
> >
> > Send Lustre-discuss mailing list submissions to
> > lustre-discuss at lists.lustre.org
> >
> > To subscribe or unsubscribe via the World Wide Web, visit
> > http://lists.lustre.org/mailman/listinfo/lustre-discuss
> > or, via email, send a message with subject or body 'help' to
> > lustre-discuss-request at lists.lustre.org
> >
> > You can reach the person managing the list at
> > lustre-discuss-owner at lists.lustre.org
> >
> > When replying, please edit your Subject line so it is more specific
> > than "Re: Contents of Lustre-discuss digest..."
> >
> >
> > Today's Topics:
> >
> > 1. Multiple NICs per OST (Joshua Bower-Cooley)
> > 2. obd error on MGS/MDT node (Jeremy Mann)
> >
> >
> > ----------------------------------------------------------------------
> >
> > Message: 1
> > Date: Tue, 26 Feb 2008 11:55:19 -0800
> > From: Joshua Bower-Cooley <jbowercooley at lcogt.net>
> > Subject: [Lustre-discuss] Multiple NICs per OST
> > To: lustre-discuss at lists.lustre.org
> > Message-ID: <200802261155.19907.jbowercooley at lcogt.net>
> > Content-Type: text/plain; charset="utf-8"
> >
> > Hi-
> > I've had lustre running happily for some time over a single 10g ethernet NIC
> > per node. After switching to dual 10g and creating a new filesystem
> > (1.6.4.2), I'm seeing nothing but keep-alive packets with bad checksums.
> >
> > What is the current "correct" way to do this now? The manual suggest not
> > using
> > bonding, but several list postings now reccommend it. Without bonding, do I
> > need to have my 2 switches stacked, or will Lustre recognize the division in
> > my subnet?
> >
> > LNET module options I've tried are:
> > 1) networks="tcp0(eth2,eth3)"
> > 2) ip2nets="tcp(eth2,eth3); tcp(eth2) 10.9.[1-4].*; tcp(eth3) 10.9.[5-8].*;"
> > and many other variations
> >
> > All other services are functional on both interfaces.
> >
> > Thanks in advance,
> > J Bower-Cooley
> > Scientific Systems Engineer
> > Las Cumbres Observatory Global Telescope
> >
> >
> > ------------------------------
> >
> > Message: 2
> > Date: Wed, 27 Feb 2008 09:50:08 -0600 (CST)
> > From: "Jeremy Mann" <jeremy at biochem.uthscsa.edu>
> > Subject: [Lustre-discuss] obd error on MGS/MDT node
> > To: lustre-discuss at lists.lustre.org
> > Message-ID:
> > <62901.24.173.62.147.1204127408.squirrel at biochem.uthscsa.edu>
> > Content-Type: text/plain;charset=iso-8859-1
> >
> > Today we are starting to see obd_change_cbdata errors on our MGS/MDT node.
> > What does this error mean?
> >
> > Version of lustre and kernel is:
> >
> > 2.6.9-42.0.10.EL_lustre-1.6.0.1smp X86_64
> >
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata()) Skipped 1
> > previous similar message
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata())
> > obd_change_cbdata: NULL export
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata()) Skipped 1
> > previous similar message
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata())
> > obd_change_cbdata: NULL export
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata()) Skipped 2
> > previous similar messages
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata())
> > obd_change_cbdata: NULL export
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata()) Skipped 4
> > previous similar messages
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata())
> > obd_change_cbdata: NULL export
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata()) Skipped 5
> > previous similar messages
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata())
> > obd_change_cbdata: NULL export
> > LustreError: 65:0:(obd_class.h:1171:obd_change_cbdata()) Skipped 1
> > previous similar message
> >
> > --
> > Jeremy Mann
> > jeremy at biochem.uthscsa.edu
> >
> > University of Texas Health Science Center
> > Bioinformatics Core Facility
> > http://www.bioinformatics.uthscsa.edu
> > Phone: (210) 567-2672
> >
> >
> >
> > ------------------------------
> >
> > _______________________________________________
> > Lustre-discuss mailing list
> > Lustre-discuss at lists.lustre.org
> > http://lists.lustre.org/mailman/listinfo/lustre-discuss
> >
> >
> > End of Lustre-discuss Digest, Vol 25, Issue 56
> > **********************************************
> >
> >
> > ------------------------------------------------------------------------
> >
> > _______________________________________________
> > Lustre-discuss mailing list
> > Lustre-discuss at lists.lustre.org
> > http://lists.lustre.org/mailman/listinfo/lustre-discuss
>
>
> --
> Amit Sharma
> Lustre Engineering
> Sun Microsystems, Bangalore.
>
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: http://lists.lustre.org/pipermail/lustre-discuss/attachments/20080229/39e94cf4/attachment.html
>
> ------------------------------
>
> _______________________________________________
> Lustre-discuss mailing list
> Lustre-discuss at lists.lustre.org
> http://lists.lustre.org/mailman/listinfo/lustre-discuss
>
>
> End of Lustre-discuss Digest, Vol 25, Issue 58
> **********************************************
>
> _______________________________________________
> Lustre-discuss mailing list
> Lustre-discuss at lists.lustre.org
> http://lists.lustre.org/mailman/listinfo/lustre-discuss
Cheers, Andreas
--
Andreas Dilger
Sr. Staff Engineer, Lustre Group
Sun Microsystems of Canada, Inc.
More information about the lustre-discuss
mailing list