Thomas Lau
2011-Mar-22 06:52 UTC
[Ocfs2-users] bug resolve yet for export OCFS2 volume to NFS client ?
I found this from ocfs1.4 document: g) NFS OCFS2 volumes can be exported as NFS volumes. This support is limited to NFS version 3, which translates to Linux kernel version 2.4 or later. Users must mount the NFS volumes on the clients using the nordirplus mount option. This disables the READDIRPLUS RPC call to workaround a bug in NFSD, detailed in the following link: http://oss.oracle.com/pipermail/ocfs2-announce/2008-June/000025.html Users running NFS version 2 can export the volume after having disabled subtree checking (mount option no_subtree_check). Be warned, disabling the check has security implications (documented in the exports(5) man page) that users must evaluate on their own. is it being fixed with centos 5.3 (2.6.18-164.11.1.el5)? anything I need to aware for exporting OCFS2 volume via NFS? -- Thomas Lau Infrastructure Delivery Manager eSun Holdings Limited Mobile: 852-93239670 Office phone: 29058104 "always I strive to push the boundaries of what we know, and what seems possible to us at this moment in time. The walls between art and engineering exist only in our minds, and few have the imagination to see beyond them." ? Theo Jansen
Sérgio Surkamp
2011-Mar-22 15:12 UTC
[Ocfs2-users] bug resolve yet for export OCFS2 volume to NFS client ?
Hi, Our NFS setup is running stable since we have disabled READDIRPLUS on our FreeBSD clients. The OCFS2 cluster servers are with about 200 days of uptime now and last reboot was to upgrade the kernel version. As a general recommendation, I suggest you to use the Oracle's Unbreakable Linux kernel on top of CentOS 5.4 (or RedHat EL 5.4) and OCFS2 1.6, also provided by Oracle, to avoid the known fragmentation bug and to be able to enable the indexed directories feature for improved performance. If you are using only Linux NFS clients, you can try running with READDIRPLUS enabled, but maybe you experience the same bug as FreeBSD NFS client and disabling the READDIRPLUS call is the only known way to avoid the problem. To install the Unbreakable, fallow the instructions provided by Oracle on its yum repository: http://public-yum.oracle.com/ Regards, S?rgio Em Tue, 22 Mar 2011 14:52:11 +0800 Thomas Lau <thomaslau at esun.com> escreveu:> I found this from ocfs1.4 document: > g) NFS > OCFS2 volumes can be exported as NFS volumes. This support is limited > to NFS version 3, which translates to Linux kernel version 2.4 or > later. Users must mount the NFS volumes on the clients using the > nordirplus mount option. This disables the READDIRPLUS RPC call to > workaround a bug in NFSD, detailed in the following link: > http://oss.oracle.com/pipermail/ocfs2-announce/2008-June/000025.html > Users running NFS version 2 can export the volume after having > disabled subtree checking (mount option no_subtree_check). Be warned, > disabling the check has security implications (documented in the > exports(5) man page) that users must evaluate on their own. > > is it being fixed with centos 5.3 (2.6.18-164.11.1.el5)? > > anything I need to aware for exporting OCFS2 volume via NFS? >-- .:''''':. .:' ` S?rgio Surkamp | Gerente de Rede :: ........ sergio at gruposinternet.com.br `:. .:' `:, ,.:' *Grupos Internet S.A.* `: :' R. Lauro Linhares, 2123 Torre B - Sala 201 : : Trindade - Florian?polis - SC :.' :: +55 48 3234-4109 : ' http://www.gruposinternet.com.br
Sérgio Surkamp
2011-Mar-22 19:22 UTC
[Ocfs2-users] bug resolve yet for export OCFS2 volume to NFS client ?
Yes. The OCFS2 mount point become unaccessible if the bug occur and the hanged OCFS2 node need to be rebooted to make it available again. The other OCFS2 nodes will block until the hanged node timeout or reboot, resuming normal operation after it. Any attempt to read or write will block the attempting process in D state. So disabling the READDIRPLUS on NFS clients is the safest option to export the OCFS2 volume and avoid/workaround the bug. Regards, S?rgio Em Wed, 23 Mar 2011 00:47:29 +0800 (GMT+08:00) Thomas Lau <thomaslau at esun.com> escreveu:> so READDIRPLUS will deadlock ocfs2 itself ?! > > ----- Original Message ----- > From: "S?rgio Surkamp" <sergio at gruposinternet.com.br> > To: "Thomas Lau" <thomaslau at esun.com> > Sent: Wednesday, March 23, 2011 12:37:55 AM > Subject: Re: [Ocfs2-users] bug resolve yet for export OCFS2 volume to > NFS client ? > > NFS READDIRPLUS bug: > http://oss.oracle.com/bugzilla/show_bug.cgi?id=977 > > Fragmentation bug: > http://oss.oracle.com/bugzilla/show_bug.cgi?id=1189 > > Regards, > S?rgio > > Em Tue, 22 Mar 2011 23:34:14 +0800 (GMT+08:00) > Thomas Lau <thomaslau at esun.com> escreveu: > > > > > what is exactly the bug is ?! > > > > > > ----- Original Message ----- > > From: "S?rgio Surkamp" <sergio at gruposinternet.com.br> > > To: ocfs2-users at oss.oracle.com > > Sent: Tuesday, March 22, 2011 11:12:59 PM > > Subject: Re: [Ocfs2-users] bug resolve yet for export OCFS2 volume > > to NFS client ? > > > > Hi, > > > > Our NFS setup is running stable since we have disabled READDIRPLUS > > on our FreeBSD clients. The OCFS2 cluster servers are with about > > 200 days of uptime now and last reboot was to upgrade the kernel > > version. > > > > As a general recommendation, I suggest you to use the Oracle's > > Unbreakable Linux kernel on top of CentOS 5.4 (or RedHat EL 5.4) and > > OCFS2 1.6, also provided by Oracle, to avoid the known fragmentation > > bug and to be able to enable the indexed directories feature for > > improved performance. > > > > If you are using only Linux NFS clients, you can try running with > > READDIRPLUS enabled, but maybe you experience the same bug as > > FreeBSD NFS client and disabling the READDIRPLUS call is the only > > known way to avoid the problem. > > > > To install the Unbreakable, fallow the instructions provided by > > Oracle on its yum repository: > > > > http://public-yum.oracle.com/ > > > > Regards, > > S?rgio > > > > Em Tue, 22 Mar 2011 14:52:11 +0800 > > Thomas Lau <thomaslau at esun.com> escreveu: > > > > > I found this from ocfs1.4 document: > > > g) NFS > > > OCFS2 volumes can be exported as NFS volumes. This support is > > > limited to NFS version 3, which translates to Linux kernel version > > > 2.4 or later. Users must mount the NFS volumes on the clients > > > using the nordirplus mount option. This disables the READDIRPLUS > > > RPC call to workaround a bug in NFSD, detailed in the following > > > link: > > > http://oss.oracle.com/pipermail/ocfs2-announce/2008-June/000025.html > > > Users running NFS version 2 can export the volume after having > > > disabled subtree checking (mount option no_subtree_check). Be > > > warned, disabling the check has security implications (documented > > > in the exports(5) man page) that users must evaluate on their own. > > > > > > is it being fixed with centos 5.3 (2.6.18-164.11.1.el5)? > > > > > > anything I need to aware for exporting OCFS2 volume via NFS? > > > > > > > > > > > >-- .:''''':. .:' ` S?rgio Surkamp | Gerente de Rede :: ........ sergio at gruposinternet.com.br `:. .:' `:, ,.:' *Grupos Internet S.A.* `: :' R. Lauro Linhares, 2123 Torre B - Sala 201 : : Trindade - Florian?polis - SC :.' :: +55 48 3234-4109 : ' http://www.gruposinternet.com.br
Sunil Mushran
2011-Mar-22 19:55 UTC
[Ocfs2-users] bug resolve yet for export OCFS2 volume to NFS client ?
On 03/21/2011 11:52 PM, Thomas Lau wrote:> I found this from ocfs1.4 document: > g) NFS > OCFS2 volumes can be exported as NFS volumes. This support is limited to > NFS version 3, which translates to Linux kernel version 2.4 or later. > Users must mount the NFS volumes on the clients using the nordirplus > mount option. This disables the READDIRPLUS RPC call to workaround a bug > in NFSD, detailed in the following link: > http://oss.oracle.com/pipermail/ocfs2-announce/2008-June/000025.html > Users running NFS version 2 can export the volume after having disabled > subtree checking (mount option no_subtree_check). Be warned, disabling > the check has security implications (documented in the exports(5) man > page) that users must evaluate on their own. > > is it being fixed with centos 5.3 (2.6.18-164.11.1.el5)? > > anything I need to aware for exporting OCFS2 volume via NFS?This is a NFS bug. It has since been fixed in mainline. Kernels based on 2.6.32+ should be ok.
Thomas Lau
2011-Mar-23 01:13 UTC
[Ocfs2-users] bug resolve yet for export OCFS2 volume to NFS client ?
is it limited to NFSv3 or NFSv4 or both NFS? out of topic question, how could I disable READDIRPLUS ? should I disable it on server side or NFS client side? ----- Original Message ----- From: "Sunil Mushran" <sunil.mushran at oracle.com> To: "Thomas Lau" <thomaslau at esun.com> Cc: ocfs2-users at oss.oracle.com Sent: Wednesday, March 23, 2011 3:55:45 AM Subject: Re: [Ocfs2-users] bug resolve yet for export OCFS2 volume to NFS client ? On 03/21/2011 11:52 PM, Thomas Lau wrote:> I found this from ocfs1.4 document: > g) NFS > OCFS2 volumes can be exported as NFS volumes. This support is limited to > NFS version 3, which translates to Linux kernel version 2.4 or later. > Users must mount the NFS volumes on the clients using the nordirplus > mount option. This disables the READDIRPLUS RPC call to workaround a bug > in NFSD, detailed in the following link: > http://oss.oracle.com/pipermail/ocfs2-announce/2008-June/000025.html > Users running NFS version 2 can export the volume after having disabled > subtree checking (mount option no_subtree_check). Be warned, disabling > the check has security implications (documented in the exports(5) man > page) that users must evaluate on their own. > > is it being fixed with centos 5.3 (2.6.18-164.11.1.el5)? > > anything I need to aware for exporting OCFS2 volume via NFS?This is a NFS bug. It has since been fixed in mainline. Kernels based on 2.6.32+ should be ok.
Sunil Mushran
2011-Mar-23 01:23 UTC
[Ocfs2-users] bug resolve yet for export OCFS2 volume to NFS client ?
Last I checked readdirplus was not in nfs4. But you may want to reconfirm. readdirplus is a rpc call initiated by the client. So has to be disabled on the nfs client. On 03/22/2011 06:13 PM, Thomas Lau wrote:> is it limited to NFSv3 or NFSv4 or both NFS? > > out of topic question, how could I disable READDIRPLUS ? > should I disable it on server side or NFS client side? > > > ----- Original Message ----- > From: "Sunil Mushran"<sunil.mushran at oracle.com> > To: "Thomas Lau"<thomaslau at esun.com> > Cc: ocfs2-users at oss.oracle.com > Sent: Wednesday, March 23, 2011 3:55:45 AM > Subject: Re: [Ocfs2-users] bug resolve yet for export OCFS2 volume to NFS client ? > > On 03/21/2011 11:52 PM, Thomas Lau wrote: >> I found this from ocfs1.4 document: >> g) NFS >> OCFS2 volumes can be exported as NFS volumes. This support is limited to >> NFS version 3, which translates to Linux kernel version 2.4 or later. >> Users must mount the NFS volumes on the clients using the nordirplus >> mount option. This disables the READDIRPLUS RPC call to workaround a bug >> in NFSD, detailed in the following link: >> http://oss.oracle.com/pipermail/ocfs2-announce/2008-June/000025.html >> Users running NFS version 2 can export the volume after having disabled >> subtree checking (mount option no_subtree_check). Be warned, disabling >> the check has security implications (documented in the exports(5) man >> page) that users must evaluate on their own. >> >> is it being fixed with centos 5.3 (2.6.18-164.11.1.el5)? >> >> anything I need to aware for exporting OCFS2 volume via NFS? > This is a NFS bug. It has since been fixed in mainline. Kernels > based on 2.6.32+ should be ok.
Thomas Lau
2011-Mar-23 01:26 UTC
[Ocfs2-users] bug resolve yet for export OCFS2 volume to NFS client ?
Ok, just in case the nfs cause ocfs2 hang, should I just powercycle that hang box and ocfs2 will resume to normal? -----Original Message----- From: Sunil Mushran Sent: 23/03/2011, 9:23 AM To: Thomas Lau Cc: ocfs2-users at oss.oracle.com Subject: Re: [Ocfs2-users] bug resolve yet for export OCFS2 volume to NFS client ? Last I checked readdirplus was not in nfs4. But you may want to reconfirm. readdirplus is a rpc call initiated by the client. So has to be disabled on the nfs client. On 03/22/2011 06:13 PM, Thomas Lau wrote:> is it limited to NFSv3 or NFSv4 or both NFS? > > out of topic question, how could I disable READDIRPLUS ? > should I disable it on server side or NFS client side? > > > ----- Original Message ----- > From: "Sunil Mushran"<sunil.mushran at oracle.com> > To: "Thomas Lau"<thomaslau at esun.com> > Cc: ocfs2-users at oss.oracle.com > Sent: Wednesday, March 23, 2011 3:55:45 AM > Subject: Re: [Ocfs2-users] bug resolve yet for export OCFS2 volume to NFS client ? > > On 03/21/2011 11:52 PM, Thomas Lau wrote: >> I found this from ocfs1.4 document: >> g) NFS >> OCFS2 volumes can be exported as NFS volumes. This support is limited to >> NFS version 3, which translates to Linux kernel version 2.4 or later. >> Users must mount the NFS volumes on the clients using the nordirplus >> mount option. This disables the READDIRPLUS RPC call to workaround a bug >> in NFSD, detailed in the following link: >> http://oss.oracle.com/pipermail/ocfs2-announce/2008-June/000025.html >> Users running NFS version 2 can export the volume after having disabled >> subtree checking (mount option no_subtree_check). Be warned, disabling >> the check has security implications (documented in the exports(5) man >> page) that users must evaluate on their own. >> >> is it being fixed with centos 5.3 (2.6.18-164.11.1.el5)? >> >> anything I need to aware for exporting OCFS2 volume via NFS? > This is a NFS bug. It has since been fixed in mainline. Kernels > based on 2.6.32+ should be ok.