pool/nfs-utils
SHA256
2
0
Fork 1
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

- nfs upstream version 1.2.5. Lots of bug fixes

Browse Source 
and improved support for pNFS.
  Adds nfsdcld daemon on osd_login script.
  Removed some very out-of-date documentation.

OBS-URL: https://build.opensuse.org/package/show/Base:System/nfs-utils?expand=0&rev=73
This commit is contained in:
Neil Brown 2012-05-17 07:07:12 +00:00 committed by Git OBS Bridge
parent bf6b9f6466
commit 92a46a2e34
9 changed files with 152 additions and 286 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

118
README.NFSv4
View File

@ -1,5 +1,5 @@
NFSv4 README
Last updated: 15 June 2006
Last updated: 17 May 2012
0. Contents:
-----------
@ -39,7 +39,7 @@ Last updated: 15 June 2006
1. Overview:
------------
The Network File System Version 4 (NFSv4) is a new distributed file system
The Network File System Version 4 (NFSv4) is a distributed file system
similar to previous versions of NFS in its straightforward design, and
independence of transport protocols and operating systems for file access in a
heterogeneous network. Unlike earlier versions of NFS, the new protocol
@ -52,7 +52,8 @@ Note: NFSv4 ACLs and krb5p (Kerberos Privacy) are currently not supported
1.1 The Purpose of this document
________________________________
This document is intended as a step-by-step guide to setup NFSv4 on SLES 10.
This document is intended as a step-by-step guide to setup NFSv4 on
openSUSE 12.
It discusses NFSv4 server and client configuration.
@ -61,14 +62,11 @@ It discusses NFSv4 server and client configuration.
For NFSv4 server:
1) Edit /etc/exports to have an entry similar to the one below:
/export <client_ip/hostname/wildcard>(rw,fsid=0,sync,no_root_squash)
(i) fsid=0 is a must.
(ii) Replace "/export" with file tree that needs to be nfs-exported and
the <client_ip/hostname/wildcard> with client's ip or hostname or *.
(* means any client)
1) /etc/exports does not require any special entries to work with
NFSv4. Earlier SUSE releases required 'fsid=0' on precisely one
entry, and 'bind=' annotations on others. This is no longer required
and should be removed. It is still supported, so there is no need
to change /etc/exports when upgrading to openSUSE 12.
2) Edit /etc/idmapd.conf to modify the default "Domain" to contain your
DNS domain name.
@ -124,88 +122,11 @@ ___________________________
There are three main configuration files you will need to edit to set up an
NFSv4 server:
/etc/exports, /etc/sysconfig/nfs and /etc/idmapd.conf.
we will describe the first two here as idmapd.conf is done in previous section.
/etc/sysconfig/nfs and /etc/idmapd.conf.
we will describe the first here as idmapd.conf is done in previous section.
4.1.1 /etc/exports
==================
This file contains a list of entries; each entry indicates a volume that is
shared and how it is shared. The /etc/exports file format is slightly
different from previous versions. A sample exports entry looks like this.
/export *(rw,fsid=0,no_subtree_check,sync,no_root_squash)
Note that:
i) fsid - The value 0 has a special meaning when use with NFSv4. NFSv4 has a
concept of a root of the overall exported filesystem. The export point
exported with fsid=0 will be used as this root.
There must be at least one entry with fsid=0. (this will be pseudo file
system's /)
ii) The method used to mount multiple exported trees is different. NFSv4 uses
the concept of pseudo filesystem to give a single file system view to the
client with a pseudo-"/" as root of the filesystem tree. To illustrate,
Suppose we have
/path1/volume1
/path2/volume2
as two filesystem trees on the server that need to be exported, then
Firstly, these need to be bound to another name under /export directory
using mount command's bind option. This is done as :
mount --bind <old dir> /export/<new dir>
i.e. in our example:
#mount --bind /path1/volume1 /export/volume1
#mount --bind /path2/volume2 /export/volume2
will bind these local filesystem trees to their local new names.
Then these two exported filesystems (with their newly bound paths) are
entered into /etc/exports with their respective independent options.
i.e. /etc/exports would contain -
/export/volume1 *(<options_to_be_filled>)
/export/volume2 *(<options_to_be_filled>)
iii)If both a directory and its subdirectory residing on different file systems
need to be exported, then the option 'nohide' must be appropriately used.
/export and /export/subdir are on differnt file systems
and both need to be exported to same client then
/export <client>(<options>)
/export/subdir <client>(<options>,nohide)
must be done so that the client can see the contents of subdir too.
Though this is not specific to NFSv4, it is seen as a common use case
scenario and is included here.
'man exports' has more info.
iv) Currently Yast2's nfs-server module can only be used as a subsitute
for manually editing the /etc/exports. Fully functional yast with other
configuration editing (idmapd etc) is work in progress.
v) In case of different kind of exports for the same exported path the
syntax that must be followed is either of the following
/export host1(<options>) host2(<options>)
(or)
/export host1(<options>)
/export host2(<options>)
4.1.2 Co-existing NFSv3 and NFSv4 exports for same file systems
===============================================================
NFSv4 current linux implementation caters to serving NFSv2 and NFSv3 clients
too. The /etc/exports can contain both type of export entries even for the
same filesystem trees being exported.
4.1.3 /etc/sysconfig/nfs
4.1.1 /etc/sysconfig/nfs
=========================
/etc/sysconfig/nfs is another NFS server configuration file. Here the number
@ -358,13 +279,12 @@ Sample configuration
Typical entries for kerberos security mode looks like these:
/export gss/krb5(rw,fsid=0,insecure,no_subtree_check,sync,no_root_squash)
/export gss/krb5i(rw,fsid=0,insecure,no_subtree_check,sync,no_root_squash)
/export gss/krb5(rw,insecure,no_subtree_check,sync,no_root_squash)
/export gss/krb5i(rw,insecure,no_subtree_check,sync,no_root_squash)
Note:
i) krb5p (Privacy) is currently not supported.
ii) option 'insecure' - The insecure option in this entry also allows clients
i) option 'insecure' - The insecure option in this entry also allows clients
with NFS implementations that don't use a reserved port for NFS. So it is
advisable *NOT* to use this option unless you have a kerberised set up or
you know what you are doing.
@ -446,14 +366,10 @@ _______________________________________________
5. showmount -e <server name>
to check mount information on NFS server
6. Make sure that one and only one path is exported
with fsid=0.
Refer Pseudofilesystems (point (iii) in Section 3.2.1) for more information.
7. If users are not mapped properly check whether idmapd is running in both
6. If users are not mapped properly check whether idmapd is running in both
server & client and dns domain name is properly configured.
8. If you unable to mount, check for the correctness of the exports file entry.
7. If you unable to mount, check for the correctness of the exports file entry.
6.2 Check list to ensure kerberos is working properly

34
mkdir-sbin Normal file
View File

@ -0,0 +1,34 @@
commit 03bb227402ab023f4badb515022d49f82e01ff8d
Author: Neil Brown <neilb@suse.de>
Date: Thu May 17 16:40:40 2012 +1000
osd_login - ensure /sbin is created before installation.
If we use a more standard approach to describing the osd_login
script, the automake infrastructure will create /sbin before
attempting installation.
This is important for: make DESTDIR=/empty-dir install
Signed-off-by: NeilBrown <neilb@suse.de>
diff --git a/utils/osd_login/Makefile.am b/utils/osd_login/Makefile.am
index adc493a..d17ffa7 100644
--- a/utils/osd_login/Makefile.am
+++ b/utils/osd_login/Makefile.am
@@ -1,12 +1,9 @@
## Process this file with automake to produce Makefile.in
-OSD_LOGIN_FILES= osd_login
+# These binaries go in /sbin (not /usr/sbin), and that cannot be
+# overriden at config time.
+sbindir = /sbin
-EXTRA_DIST= $(OSD_LOGIN_FILES)
-
-all-local: $(OSD_LOGIN_FILES)
-
-install-data-hook:
- $(INSTALL) --mode 755 osd_login $(DESTDIR)/sbin/osd_login
+sbin_SCRIPTS = osd_login
MAINTAINERCLEANFILES = Makefile.in

3
nfs-utils-1.2.5.tar.bz2
View File

@ -1,3 +0,0 @@
version https://git-lfs.github.com/spec/v1
oid sha256:3595ed4f6ee5c13514540861ad5e89907aa4fd6897bfa99577515233e4745a9b
size 698148

3
nfs-utils-1.2.6.tar.bz2 Normal file
View File

@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:70267288500408495085b7c54a667c4e8cfde5882f3193da095fc7dad2da0b83
size 755305

11
nfs-utils-page_size.patch
View File

@ -1,11 +0,0 @@
--- utils/blkmapd/device-process.c.orig
+++ utils/blkmapd/device-process.c
@@ -296,7 +296,7 @@ decode_blk_volume(uint32_t **pp, uint32_
off_t stripe_unit = vol->param.bv_stripe_unit;
/* Check limitations imposed by device-mapper */
if ((stripe_unit & (stripe_unit - 1)) != 0
- || stripe_unit < (off_t) (PAGE_SIZE >> 9))
+ || stripe_unit < (off_t) (sysconf(_SC_PAGESIZE) >> 9))
return -EIO;
BLK_READBUF(p, end, 4);
READ32(vol->bv_vol_n);

8
nfs-utils.changes
View File

@ -1,3 +1,11 @@
-------------------------------------------------------------------
Thu May 17 07:04:58 UTC 2012 - nfbrown@suse.com
- nfs upstream version 1.2.5. Lots of bug fixes
and improved support for pNFS.
Adds nfsdcld daemon on osd_login script.
Removed some very out-of-date documentation.
-------------------------------------------------------------------
Wed May 2 01:42:23 UTC 2012 - nfbrown@suse.com

22
nfs-utils.spec
View File

@ -21,6 +21,7 @@ BuildRequires: device-mapper-devel
BuildRequires: e2fsprogs-devel
BuildRequires: gcc-c++
BuildRequires: krb5-devel
BuildRequires: libevent-devel
BuildRequires: libgssglue-devel >= 0.3
BuildRequires: libmount-devel
BuildRequires: librpcsecgss
@ -28,17 +29,13 @@ BuildRequires: libtirpc-devel
BuildRequires: libtool
BuildRequires: nfsidmap-devel >= 0.24
BuildRequires: pkgconfig
BuildRequires: sqlite3-devel
BuildRequires: tcpd-devel
%if 0%{?suse_version} > 1100
BuildRequires: libevent-devel
%else
BuildRequires: libevent
%endif
Url: http://nfs.sourceforge.net
Summary: Support Utilities for Kernel nfsd
License: GPL-2.0+
Group: Productivity/Networking/NFS
Version: 1.2.5
Version: 1.2.6
Release: 0
BuildRoot: %{_tmppath}/%{name}-%{version}-build
PreReq: %fillup_prereq %insserv_prereq
@ -60,8 +57,8 @@ Source11: idmapd.conf
Source12: start-statd
Source13: nfs-utils.rpmlintrc
Patch0: nfs-utils-1.0.7-bind-syntax.patch
Patch1: warn-nfs-udp.patch
Patch2: nfs-utils-page_size.patch
Patch1: remove_pretty_sig.patch
Patch2: mkdir-sbin
Suggests: python-base
%description
@ -139,11 +136,10 @@ Authors:
%setup -q -n nfs-utils-%{version} -a 1
%patch0 -p1
%patch1 -p1
%patch2
%patch2 -p1
cp %{S:6} .
%build
%{?suse_update_config:%{suse_update_config -f }}
rm -f configure; autoreconf -fi
CFLAGS="$RPM_OPT_FLAGS -fPIE -fno-strict-aliasing" LDFLAGS="-pie" ./configure \
--mandir=%{_mandir} \
@ -151,6 +147,7 @@ CFLAGS="$RPM_OPT_FLAGS -fPIE -fno-strict-aliasing" LDFLAGS="-pie" ./configure \
--enable-nfsv4 \
--enable-gss \
--enable-ipv6 \
--enable-nfsdcld \
--enable-mount \
--enable-libmount-mount \
--enable-mountconfig \
@ -266,6 +263,7 @@ fi
/sbin/mount.nfs4
/sbin/umount.nfs
/sbin/umount.nfs4
/sbin/osd_login
/usr/sbin/gss_clnt_send_err
/usr/sbin/gss_destroy_creds
%attr(0755,root,root) /usr/sbin/mountstats
@ -318,6 +316,7 @@ fi
/usr/sbin/rpc.mountd
/usr/sbin/rpc.nfsd
/usr/sbin/rpc.svcgssd
/usr/sbin/nfsdcld
%{_mandir}/man5/exports.5.gz
%{_mandir}/man7/nfsd.7.gz
%{_mandir}/man8/exportfs.8.gz
@ -326,6 +325,7 @@ fi
%{_mandir}/man8/rpc.mountd.8.gz
%{_mandir}/man8/rpc.nfsd.8.gz
%{_mandir}/man8/rpc.svcgssd.8.gz
%{_mandir}/man8/nfsdcld.8.gz
/usr/share/omc/svcinfo.d/nfs-kernel-server.xml
%config(noreplace) /var/lib/nfs/xtab
%config(noreplace) /var/lib/nfs/etab
@ -334,6 +334,6 @@ fi
%files -n nfs-doc
%defattr(-,root,root)
%doc nfs/*.html nfs/*.ps linux-nfs/* README.NFSv4
%doc nfs/*.html nfs/*.ps README.NFSv4
%changelog

79
remove_pretty_sig.patch Normal file
View File

@ -0,0 +1,79 @@
From: Jim Rees <rees@umich.edu>
Date: Tue, 27 Mar 2012 10:18:39 -0400
Subject: [PATCH 1/2] remove pretty_sig()
This really only works against EMC servers. There is at least one server
that returns unprintable signatures, which fill the log with garbage (the
spec does not mandate printable signatures). It could be made more generic,
for example by checking each byte for isprint(). But the signatures are
really only of interest to developers, not admins, so it seems better to
just remove it.
Signed-off-by: Jim Rees <rees@umich.edu>
Acked-by: NeilBrown <neilb@suse.de>
---
utils/blkmapd/device-process.c | 29 -----------------------------
1 file changed, 29 deletions(-)
diff --git a/utils/blkmapd/device-process.c b/utils/blkmapd/device-process.c
index 652a7a8..5fe3dff 100644
--- a/utils/blkmapd/device-process.c
+++ b/utils/blkmapd/device-process.c
@@ -49,28 +49,6 @@
#include "device-discovery.h"
-static char *pretty_sig(char *sig, uint32_t siglen)
-{
- static char rs[100];
- uint64_t sigval;
- unsigned int i;
-
- if (siglen <= sizeof(sigval)) {
- sigval = 0;
- for (i = 0; i < siglen; i++)
- sigval |= ((unsigned char *)sig)[i] << (i * 8);
- sprintf(rs, "0x%0llx", (unsigned long long) sigval);
- } else {
- if (siglen > sizeof rs - 4) {
- siglen = sizeof rs - 4;
- sprintf(&rs[siglen], "...");
- } else
- rs[siglen] = '\0';
- memcpy(rs, sig, siglen);
- }
- return rs;
-}
-
uint32_t *blk_overflow(uint32_t * p, uint32_t * end, size_t nbytes)
{
uint32_t *q = p + ((nbytes + 3) >> 2);
@@ -109,9 +87,6 @@ static int decode_blk_signature(uint32_t **pp, uint32_t * end,
* for mapping, then thrown away.
*/
comp->bs_string = (char *)p;
- BL_LOG_INFO("%s: si_comps[%d]: bs_length %d, bs_string %s\n",
- __func__, i, siglen,
- pretty_sig(comp->bs_string, siglen));
p += ((siglen + 3) >> 2);
}
*pp = p;
@@ -152,10 +127,6 @@ read_cmp_blk_sig(struct bl_disk *disk, int fd, struct bl_sig_comp *comp)
}
ret = memcmp(sig, comp->bs_string, siglen);
- if (!ret)
- BL_LOG_INFO("%s: %s sig %s at %lld\n", __func__, dev_name,
- pretty_sig(sig, siglen),
- (long long)comp->bs_offset);
out:
if (sig)
--
1.7.9.5
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

160
warn-nfs-udp.patch
View File

@ -1,160 +0,0 @@
---
utils/mount/nfs.man | 83 +++++++++++++++++++++++++++++++++++++++++++++++++
utils/mount/nfsmount.c | 3 +
utils/mount/stropts.c | 13 +++++++
3 files changed, 99 insertions(+)
--- nfs-utils-1.2.4.orig/utils/mount/nfs.man
+++ nfs-utils-1.2.4/utils/mount/nfs.man
@@ -503,6 +503,8 @@ Specifying a netid that uses TCP forces
command and the NFS client to use TCP.
Specifying a netid that uses UDP forces all traffic types to use UDP.
.IP
+.B Before using NFS over UDP, please refer to the section WARNINGS below.
+.IP
If the
.B proto
mount option is not specified, the
@@ -517,6 +519,8 @@ The
option is an alternative to specifying
.BR proto=udp.
It is included for compatibility with other operating systems.
+.IP
+.B Before using NFS over UDP, please refer to the section WARNINGS below.
.TP 1.5i
.B tcp
The
@@ -975,6 +979,8 @@ in a single frame) is advised. This r
the loss of a single MTU-sized network frame results in the loss of
an entire large read or write request.
.P
+Please see also the WARNINGS section below.
+.P
TCP is the default transport protocol used for all modern NFS
implementations. It performs well in almost every conceivable
network environment and provides excellent guarantees against data
@@ -1566,6 +1572,83 @@ export pathname, but not both, during a
merges the mount option
.B ro
with the mount options already saved on disk for the NFS server mounted at /mnt.
+.SH WARNINGS
+Using NFS over UDP on high-speed links such as Gigabit
+.BR "can cause silent data corruption" .
+.P
+The problem can be triggered at high loads, and is caused by problems in
+IP fragment reassembly. NFS read and writes typically transmit UDP packets
+of 4 Kilobytes or more, which have to be broken up into several fragments
+in order to be sent over the Ethernet link, which limits packets to 1500
+bytes by default. This process happens at the IP network layer and is
+called fragmentation.
+.P
+In order to identify fragments that belong together, IP assigns a 16bit
+.I IP ID
+value to each packet; fragments generated from the same UDP packet
+will have the same IP ID. The receiving system will collect these
+fragments and combine them to form the original UDP packet. This process
+is called reassembly. The default timeout for packet reassembly is
+30 seconds; if the network stack does not receive all fragments of
+a given packet within this interval, it assumes the missing fragment(s)
+got lost and discards those it already received.
+.P
+The problem this creates over high-speed links is that it is possible
+to send more than 65536 packets within 30 seconds. In fact, with
+heavy NFS traffic one can observe that the IP IDs repeat after about
+5 seconds.
+.P
+This has serious effects on reassembly: if one fragment gets lost,
+another fragment
+.I from a different packet
+but with the
+.I same IP ID
+will arrive within the 30 second timeout, and the network stack will
+combine these fragments to form a new packet. Most of the time, network
+layers above IP will detect this mismatched reassembly - in the case
+of UDP, the UDP checksum, which is a 16 bit checksum over the entire
+packet payload, will usually not match, and UDP will discard the
+bad packet.
+.P
+However, the UDP checksum is 16 bit only, so there is a chance of 1 in
+65536 that it will match even if the packet payload is completely
+random (which very often isn't the case). If that is the case,
+silent data corruption will occur.
+.P
+This potential should be taken seriously, at least on Gigabit
+Ethernet.
+Network speeds of 100Mbit/s should be considered less
+problematic, because with most traffic patterns IP ID wrap around
+will take much longer than 30 seconds.
+.P
+It is therefore strongly recommended to use
+.BR "NFS over TCP where possible" ,
+since TCP does not perform fragmentation.
+.P
+If you absolutely have to use NFS over UDP over Gigabit Ethernet,
+some steps can be taken to mitigate the problem and reduce the
+probability of corruption:
+.TP +1.5i
+.I Jumbo frames:
+Many Gigabit network cards are capable of transmitting
+frames bigger than the 1500 byte limit of traditional Ethernet, typically
+9000 bytes. Using jumbo frames of 9000 bytes will allow you to run NFS over
+UDP at a page size of 8K without fragmentation. Of course, this is
+only feasible if all involved stations support jumbo frames.
+.IP
+To enable a machine to send jumbo frames on cards that support it,
+it is sufficient to configure the interface for a MTU value of 9000.
+.TP +1.5i
+.I Lower reassembly timeout:
+By lowering this timeout below the time it takes the IP ID counter
+to wrap around, incorrect reassembly of fragments can be prevented
+as well. To do so, simply write the new timeout value (in seconds)
+to the file
+.BR /proc/sys/net/ipv4/ipfrag_time .
+.IP
+A value of 2 seconds will greatly reduce the probability of IPID clashes on
+a single Gigabit link, while still allowing for a reasonable timeout
+when receiving fragmented traffic from distant peers.
.SH FILES
.TP 1.5i
.I /etc/fstab
--- nfs-utils-1.2.4.orig/utils/mount/nfsmount.c
+++ nfs-utils-1.2.4/utils/mount/nfsmount.c
@@ -264,6 +264,9 @@ parse_options(char *old_opts, struct nfs
if (!strcmp(opteq+1, "udp")) {
nfs_pmap->pm_prot = IPPROTO_UDP;
mnt_pmap->pm_prot = IPPROTO_UDP;
+ fprintf(stderr,
+ "Using NFS over UDP can cause data corruption.\n"
+ "Please refer to the WARNINGS section of the nfs(5) manual page.\n");
#if NFS_MOUNT_VERSION >= 2
data->flags &= ~NFS_MOUNT_TCP;
} else if (!strcmp(opteq+1, "tcp") &&
--- nfs-utils-1.2.4.orig/utils/mount/stropts.c
+++ nfs-utils-1.2.4/utils/mount/stropts.c
@@ -567,6 +567,8 @@ static int nfs_sys_mount(struct nfsmount
{
char *options = NULL;
int result;
+ char *proto;
+ static int once = 0;
if (mi->fake)
return 1;
@@ -575,6 +577,17 @@ static int nfs_sys_mount(struct nfsmount
errno = EIO;
return 0;
}
+ if (po_contains(mi->options, "udp"))
+ proto = "udp";
+ else
+ proto = po_get(mi->options, "proto");
+ if (proto && strcmp(proto, "udp") == 0 && !once) {
+ fprintf(stderr,
+ "Using NFS over UDP can cause data corruption.\n"
+ "Please refer to the WARNINGS section of the nfs(5) manual page.\n");
+ once=1;
+ }
+
result = mount(mi->spec, mi->node, mi->type,
mi->flags & ~(MS_USER|MS_USERS), options);