IPv6 support
Subsystems (2)
Hi Chuck et al,
Thanks for all the excellent work on IPv6 support. There have been considerable amount of development that has been happening recently. However, it's not very clear what are the missing pieces or open problems (if any).
Quoted 1 line
From what I understand, the current status is (CMIW):
+ kernel support
- client side and server (kernel pieces) are almost there and the missing pieces are expected to get in by 2.6.32.
+ userspace support
- nfsd support has already been added but disabled temporarily till we get statd and mountd IPv6 aware.
- making statd IPv6 aware is currently under development/review.
- mount.nfs changes - done?
- rpc.mountd, exportfs changes - TBD.
Is the above information correct? How far we are from getting full IPv6 support?
Thanks,
Signature · 1 line
Suresh Jayaraman
Quoted 6 lines · Suresh Jayaraman
Hi Chuck et al,
Thanks for all the excellent work on IPv6 support. There have been considerable amount of development that has been happening recently. However, it's not very clear what are the missing pieces or open problems (if any).
There's a working document on the linux-nfs.org wiki that has a full list, but I can't seem to get to linux-nfs.org right now.
Quoted 5 lines
From what I understand, the current status is (CMIW):
+ kernel support
- client side and server (kernel pieces) are almost there and the missing pieces are expected to get in by 2.6.32.
The client-side implementation of IPv6 in the kernel is usable. There's been a request for the proto= mount option to accept netids instead of protocol names, and that's still pending. Otherwise, you can pass an IPv6 address to the kernel and it will do the right thing. NFSv4 is pretty much working now, NFSv2/v3 still needs the IPv6-aware statd. We've tested against NetApp and Solaris NFS servers.
Server side in the kernel is nearly done. There are a handful of small patches outstanding. Until I have mountd working, there's no way to test the server side pieces thoroughly. So for now, "completion" of the server side is some time in the future.
Quoted 4 lines
+ userspace support
- nfsd support has already been added but disabled temporarily till we get statd and mountd IPv6 aware.
- making statd IPv6 aware is currently under development/review.
Basic support is code complete and has seen some testing. We're waiting for upstream NFS maintainers to review the latest patch series.
Quoted 1 line
- mount.nfs changes - done?
The netid changes mentioned above will require some mount.nfs support, which is in the works. 2.6.33, maybe?
Quoted 1 line
- rpc.mountd, exportfs changes - TBD.
These are the next step. This is going to be a fairly heavy-weight change -- mountd and exportfs internals are unsuited for internationalized domain names, IPv6, and mixed IPv4/IPv6 environments, and will need to be replaced.
There is also the gssd twins. Jeff has fixed up at least one of them. I think we've reviewed the kerberos libraries and determined that they are IPv6-aware, but that could be incorrect. TCP wrappers are definitely not IPv6-aware.
rpcbind and libtirpc are also prerequisites, but there are few distributions I know of that have replaced portmapper with rpcbind.
Quoted 1 line
How far we are from getting full IPv6 support?
The short answer is mountd/exportfs will take about six months. The client side is close enough that I think we can promise basic IPv6 support for upcoming enterprise releases. We're also thinking the kernel work on the server is close enough that few or no kABI changes will be needed once mountd is working, but that's just a guess.
The long answer depends on your definition of "full". :-)
Does that include complete support for link-local IPv6 addresses? Does that include full support for netids in the kernel? Does that include complete multi-homed host support in lockd and statd? Does that include full support for internationalized domain names? Does that include IPv6 netgroup support? Does it include IPv6 support in TCP wrappers? Does it include support for systems that have no IPv4 addresses (not even loopback)?
There are a bunch of details that still need to be worked through. I've only been able to guess at what features are required, and which can be implemented at a later time. What I would dearly love to have is a list of specific features that folks feel is a baseline (based on actual data, of course). Unfortunately there are very few Linux customers who have deployed IPv6 and an IPv6-capable NFS implementation (like Solaris) and can tell us what they would need on Linux.
Plus, most distributions don't have fluent user space infrastructure for IPv6 yet. NetworkManager is one area that may need work. The Network Administration tool in Fedora is still IPv4-centric, iirc. We don't have firewall admin tools that handle IPv6 rules. Unlike IPv4, admins can (and often do) use IPv6-aware kernels without ipv6.ko, so all of our tools and support have to be careful about using IPv6 when the O/S may not support it. This is different than IPv4, which is nearly always available.
In other words, system support for IPv6 outside of NFS is kind of like wireless was 5 years ago -- innumerable small admin tools that users had to integrate themselves, usually without much success.
--
Chuck Lever
chuck[dot]lever[at]oracle[dot]com
Quoted 6 lines · On Mon, 2 Nov 2009 11:40:47 -0500 Chuck Lever <[email protected]> wrote:
There is also the gssd twins. Jeff has fixed up at least one of them. I think we've reviewed the kerberos libraries and determined that they are IPv6-aware, but that could be incorrect. TCP wrappers are definitely not IPv6-aware.
rpc.gssd should be ready to go for IPv6. rpc.svcgssd didn't need any changes since it only talks to the kernel.
Signature · 1 line
Jeff Layton <[email protected]>
Hi Chuck,
Thanks much for the detailed update.
Quoted 10 lines · Chuck Lever
Quoted 6 lines · Suresh Jayaraman
Hi Chuck et al,
Thanks for all the excellent work on IPv6 support. There have been considerable amount of development that has been happening recently. However, it's not very clear what are the missing pieces or open problems (if any).
There's a working document on the linux-nfs.org wiki that has a full list, but I can't seem to get to linux-nfs.org right now.
Is this what you mentioned?
http://wiki.linux-nfs.org/wiki/index.php/Server_IPv6_support
http://wiki.linux-nfs.org/wiki/index.php/NFSv4_Introduction#IPv6_support_for_the_client
Quoted 11 lines
The long answer depends on your definition of "full". :-)
Does that include complete support for link-local IPv6 addresses? Does that include full support for netids in the kernel? Does that include complete multi-homed host support in lockd and statd? Does that include full support for internationalized domain names? Does that include IPv6 netgroup support? Does it include IPv6 support in TCP wrappers? Does it include support for systems that have no IPv4 addresses (not even loopback)?
Quoted 5 lines
There are a bunch of details that still need to be worked through. I've only been able to guess at what features are required, and which can be implemented at a later time. What I would dearly love to have is a list of specific features that folks feel is a baseline (based on actual data, of course).
I'd try and see if I could get the list of specific features.
Quoted 10 lines
Plus, most distributions don't have fluent user space infrastructure for IPv6 yet. NetworkManager is one area that may need work. The Network Administration tool in Fedora is still IPv4-centric, iirc. We don't have firewall admin tools that handle IPv6 rules. Unlike IPv4, admins can (and often do) use IPv6-aware kernels without ipv6.ko, so all of our tools and support have to be careful about using IPv6 when the O/S may not support it. This is different than IPv4, which is nearly always available.
Thanks,
Signature · 1 line
Suresh Jayaraman
Quoted 13 lines · Suresh Jayaraman
Hi Chuck,
Thanks much for the detailed update.
Quoted 8 lines · Chuck Lever
Quoted 6 lines · Suresh Jayaraman
Hi Chuck et al,
Thanks for all the excellent work on IPv6 support. There have been considerable amount of development that has been happening recently. However, it's not very clear what are the missing pieces or open problems (if any).
Quoted 7 lines
Quoted 2 lines
There's a working document on the linux-nfs.org wiki that has a full list, but I can't seem to get to linux-nfs.org right now.
Is this what you mentioned?
http://wiki.linux-nfs.org/wiki/index.php/Server_IPv6_support
http://wiki.linux-nfs.org/wiki/index.php/NFSv4_Introduction#IPv6_support_for_the_client
Those are old design documents. I was referring to:
http://wiki.linux-nfs.org/wiki/index.php/Ipv6PlanningDocument
which is usually updated at least once a quarter.
--
Chuck Lever
chuck[dot]lever[at]oracle[dot]com
IPv6 support: dreaming of...
In reply to Chuck LeverQuoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
"CL" == Chuck Lever <chuck.lever...> writes:
"SJ" == Suresh Jayaraman wrote:
... And first of all: it seems that there're an immense amount of work being done to add the NFS over IPv6 support for GNU/Linux. Thanks for that!
[...]
SJ> How far we are from getting full IPv6 support?
CL> The short answer is mountd/exportfs will take about six months. CL> The client side is close enough that I think we can promise basic CL> IPv6 support for upcoming enterprise releases. We're also thinking CL> the kernel work on the server is close enough that few or no kABI CL> changes will be needed once mountd is working, but that's just a CL> guess.
CL> The long answer depends on your definition of "full". :-)
Well, I'll try to answer to it based on my own use of NFS (mostly NFSv3, at this moment), as was suggested.
Basically, every now and then I have to set up some shared filesystem-like resources. To this end, I usually deploy (at the same time):
* HTTP, for its ubiquity;
Alias /public /var/public
<Directory /var/public>
Options Indexes SymLinksIfOwnerMatch IncludesNoExec ExecCGI AllowOverride FileInfo AuthConfig Limit
</Directory>
* Rsync, for its superior handling of replication;
[public]
path = /var/public
comment = Public directory
* NFS, for it requires no specific support from the applications themselves.
/var/public *(ro,all_squash,async,insecure)/var/public/debian *(ro,all_squash,async,insecure)/var/public/storage *(ro,all_squash,async,insecure)
Mostly, these are used to conserve disk space by allowing a few copies of certain files to be used by a number of users. Rarely I need to maintain anything writable, and if possible, prefer to deploy SSH/SFTP for ``uploading''.
I need almost no access management, as most of the data was received from open sources (say, http://www.debian.org/,
http://www.landcover.org/ or http://wist.echo.nasa.gov/), and there were no DoS-doers or bandwidth hogs so far.
Actually, even NFSv3 suits my needs almost completely, though occasionally I need Kerberos.
CL> Does that include complete support for link-local IPv6 addresses?
No. Not for me, at the least. IIUC, these are to be used mainly for low-level network services and testing. It could be handy to support these widely, but I'd not make it a priority task.
CL> Does that include full support for netids in the kernel?
No clue on this, sorry.
CL> Does that include complete multi-homed host support in lockd and CL> statd?
I believe that multi-homing is going to be more likely in IPv6 environments. I've never experimented with NFS on multi-homed IPv4 hosts, though, so I'm not sure what's meant by ``complete support'' here.
CL> Does that include full support for internationalized domain names?
Not for me.
CL> Does that include IPv6 netgroup support?
Do you mean netgroup like in NIS (or LDAP with a NIS-like schema)? These also could be handy, but not a priority task, at least for me.
CL> Does it include IPv6 support in TCP wrappers?
In general, I see TCP wrappers as, roughly, a functional equivalent of netfilter / ip6tables(8). So, once again, not a priority for me. (If I don't miss something.)
CL> Does it include support for systems that have no IPv4 addresses CL> (not even loopback)?
Absolutely. But it surely can wait.
CL> There are a bunch of details that still need to be worked through. CL> I've only been able to guess at what features are required, and CL> which can be implemented at a later time. What I would dearly love CL> to have is a list of specific features that folks feel is a CL> baseline (based on actual data, of course).
Hopefully, the bits above will help.
CL> Unfortunately there are very few Linux customers who have deployed CL> IPv6 and an IPv6-capable NFS implementation (like Solaris) and can CL> tell us what they would need on Linux.
It seems that my brother's going to do me a favor and install OpenSolaris for me (which, as I saw today, is being able to export and mount filesystems using NFS over IPv6.)
(I wonder, are there Apache and Rsync bundled with OpenSolaris?)
Perhaps I'll be able to experiment with some (up to 20 or so) Debian GNU/Linux hosts using it as a file server.
CL> Plus, most distributions don't have fluent user space CL> infrastructure for IPv6 yet. NetworkManager is one area that may CL> need work. The Network Administration tool in Fedora is still CL> IPv4-centric, iirc.
... And, unfortunately, even interfaces(5) handling in Debian has flaws in that respect. Sigh.
CL> We don't have firewall admin tools that handle IPv6 rules.
Personally, I prefer ip6tables(8), which does.
CL> Unlike IPv4, admins can (and often do) use IPv6-aware kernels CL> without ipv6.ko, so all of our tools and support have to be careful CL> about using IPv6 when the O/S may not support it. This is CL> different than IPv4, which is nearly always available.
Agreed.
CL> In other words, system support for IPv6 outside of NFS is kind of CL> like wireless was 5 years ago -- innumerable small admin tools that CL> users had to integrate themselves, usually without much success.
Signature · 1 line
FSF associate member #7257
Re: IPv6 support: dreaming of...
In reply to a message not in the archiveThanks, Ivan. That echoes my own concept of a basic NFS over IPv6 deployment use case.
Quoted 140 lines · Ivan Shmakov
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
"CL" == Chuck Lever <chuck.lever...> writes:
"SJ" == Suresh Jayaraman wrote:... And first of all: it seems that there're an immense amount of work being done to add the NFS over IPv6 support for GNU/Linux. Thanks for that!
[...]
SJ> How far we are from getting full IPv6 support?
CL> The short answer is mountd/exportfs will take about six months. CL> The client side is close enough that I think we can promise basic CL> IPv6 support for upcoming enterprise releases. We're also thinking CL> the kernel work on the server is close enough that few or no kABI CL> changes will be needed once mountd is working, but that's just a CL> guess.
CL> The long answer depends on your definition of "full". :-)
Well, I'll try to answer to it based on my own use of NFS (mostly NFSv3, at this moment), as was suggested.
Basically, every now and then I have to set up some shared filesystem-like resources. To this end, I usually deploy (at the same time):
* HTTP, for its ubiquity;
Alias /public
/var/public<Directory
/var/public>
Options Indexes SymLinksIfOwnerMatch IncludesNoExec ExecCGI AllowOverride FileInfo AuthConfig Limit
</Directory>* Rsync, for its superior handling of replication;
[public]
path =/var/public
comment = Public directory* NFS, for it requires no specific support from the applications themselves.
/var/public*(ro,all_squash,async,insecure)/var/public/debian*(ro,all_squash,async,insecure)/var/public/storage*(ro,all_squash,async,insecure)Mostly, these are used to conserve disk space by allowing a few copies of certain files to be used by a number of users. Rarely I need to maintain anything writable, and if possible, prefer to deploy SSH/SFTP for ``uploading''.
I need almost no access management, as most of the data was received from open sources (say, http://www.debian.org/,
http://www.landcover.org/ or http://wist.echo.nasa.gov/), and there were no DoS-doers or bandwidth hogs so far.Actually, even NFSv3 suits my needs almost completely, though occasionally I need Kerberos.
CL> Does that include complete support for link-local IPv6 addresses?
No. Not for me, at the least. IIUC, these are to be used mainly for low-level network services and testing. It could be handy to support these widely, but I'd not make it a priority task.
CL> Does that include full support for netids in the kernel?
No clue on this, sorry.
CL> Does that include complete multi-homed host support in lockd and CL> statd?
I believe that multi-homing is going to be more likely in IPv6 environments. I've never experimented with NFS on multi-homed IPv4 hosts, though, so I'm not sure what's meant by ``complete support'' here.
CL> Does that include full support for internationalized domain names?
Not for me.
CL> Does that include IPv6 netgroup support?
Do you mean netgroup like in NIS (or LDAP with a NIS-like schema)? These also could be handy, but not a priority task, at least for me.
CL> Does it include IPv6 support in TCP wrappers?
In general, I see TCP wrappers as, roughly, a functional equivalent of netfilter / ip6tables(8). So, once again, not a priority for me. (If I don't miss something.)
CL> Does it include support for systems that have no IPv4 addresses CL> (not even loopback)?
Absolutely. But it surely can wait.
CL> There are a bunch of details that still need to be worked through. CL> I've only been able to guess at what features are required, and CL> which can be implemented at a later time. What I would dearly love CL> to have is a list of specific features that folks feel is a CL> baseline (based on actual data, of course).
Hopefully, the bits above will help.
CL> Unfortunately there are very few Linux customers who have deployed CL> IPv6 and an IPv6-capable NFS implementation (like Solaris) and can CL> tell us what they would need on Linux.
It seems that my brother's going to do me a favor and install OpenSolaris for me (which, as I saw today, is being able to export and mount filesystems using NFS over IPv6.)
(I wonder, are there Apache and Rsync bundled with OpenSolaris?)
Perhaps I'll be able to experiment with some (up to 20 or so) Debian GNU/Linux hosts using it as a file server.
CL> Plus, most distributions don't have fluent user space CL> infrastructure for IPv6 yet. NetworkManager is one area that may CL> need work. The Network Administration tool in Fedora is still CL> IPv4-centric, iirc.
... And, unfortunately, even interfaces(5) handling in Debian has flaws in that respect. Sigh.
CL> We don't have firewall admin tools that handle IPv6 rules.
Personally, I prefer ip6tables(8), which does.
CL> Unlike IPv4, admins can (and often do) use IPv6-aware kernels CL> without ipv6.ko, so all of our tools and support have to be careful CL> about using IPv6 when the O/S may not support it. This is CL> different than IPv4, which is nearly always available.
Quoted 13 lines
Agreed.
CL> In other words, system support for IPv6 outside of NFS is kind of CL> like wireless was 5 years ago -- innumerable small admin tools that CL> users had to integrate themselves, usually without much success.
Signature · 6 lines
FSF associate member #7257 -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
--
Chuck Lever
chuck[dot]lever[at]oracle[dot]com
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"CL" == Chuck Lever <chuck.lever...> writes:
[...]
CL> In other words, system support for IPv6 outside of NFS is kind of CL> like wireless was 5 years ago -- innumerable small admin tools that CL> users had to integrate themselves, usually without much success.
BTW, I tend to disagree with that. I've ``discovered'' IPv6 only this summer, and for me, the whole thing was smooth...
... Almost. There were only a few issues:
* /etc/network/interfaces (like in Debian GNU/Linux) handling is a bit IPv4-centric (even more so in Etch, oldstable); say, one may have to resort to the hacks like:
iface eth1 inet6 static
address X:Y:Z:W::1234:5678
netmask 64
gateway X:Y:Z:W::1
post-up ip -6 addr add G:H:I:J::DEAD:1/64 dev "$IFACE" && ip -6 addr add K:L:M:N::1337:CAFE/64 dev "$IFACE"
(with Etch, one'd have to add IPv6 management commands to the existing IPv4 interfaces' stanzas.)
* the stable version of Squid lacks IPv6 support completely; Apache has it, but it lacks caching, which is a must for me;
* there're no VoIP switches with IPv6 support in Debian; in particular, Asterisk speaks no IPv6, while FreeSWITCH is not in Debian; (there was an ITP); there's Linphone for an IPv6-compatible user agent;
* NFS.
Otherwise, I use IPv6 on a daily basis, both at home and at work, and there seems to be no significant flaws, other than those few above.
Occasionally, some hardware seems to behave strangely (e. g., I was unable to use IPv6 in one of classrooms here), and sometimes there're noticeable response times from hosts using old NICs, but hey, there're still a couple of 10 Mbit/s hubs in the darker corners of the network, and nobody seems to worry about it much.
Signature · 1 line
FSF associate member #7257