Revision as of 14:32, 28 January 2015

NFS bug study

All debian used are fresh install of wheezy 7.8.
All alpine used are fresh install of edge. (will also try vanilla kernel in KVM)
All boxes are supermicro servers with bi-Xeon running AL from USB key.
I do not have physical access to the boxes!

The NFS-servers are configured to export

/srv/home       192.168.1.0/24(rw,sync,no_subtree_check)

The nfs-clients are configured to mount from fstab

storage:/srv/home /home nfs noauto,defaults,noexec 0 0

"storage" is defined in /etc/hosts to point to the right server.

The test is done with

mount /home

We will compare the dmesg outputs, the ls -ld /home outputs, the cat /home/test and touch /home/toto ones. /home/test is prepared on the server (just a text file containing "do you see me?"). Those tests are run as root user.

Will redo usage tests with non root user because of the default squashroot of NFS...

NFS-server in KVM-Debian

fresh install with tasksel "file server"
this KVM in running on bare metal alpine

nfs-client in KVM AL

mount /home gives
in dmesg

[73460.112383] RPC: Registered named UNIX socket transport module.
[73460.112386] RPC: Registered udp transport module.
[73460.112388] RPC: Registered tcp transport module.
[73460.112389] RPC: Registered tcp NFSv4.1 backchannel transport module.
[73460.165060] svc: failed to register lockdv1 RPC service (errno 111).
[73460.165069] lockd_up: makesock failed, error=-111
[73460.217513] NFS: Registering the id_resolver key type
[73460.217524] Key type id_resolver registered
[73460.217525] Key type id_legacy registered

in ls -ld /home/

drwxr-xr-x    2 42949672 42949672         6 Jan 23 12:27 /home

in cat /home/test

 Do you see me?

in touch /home/toto

touch: /home/toto: Permission denied

nfs-client in KVM debian

dmesg is empty
ls -ld /home

drwxr-xr-x 2 root root 17 Jan 23 08:39 /home

cat /home/test

Do you see me?

touch /home/toto (even after adding rw to the mount options in fstab)

touch: cannot touch `/home/toto': Permission denied

Some pointers to investigate this permission problem:

http://unix.stackexchange.com/questions/79172/nfs-permission-denied

To begin using machine as an NFS client, you will need the portmapper running on that machine, and to use NFS file locking, you will also need rpc.statd and rpc.lockd running on both the client and the server.

'

nfs-client in LXC AL (on bare metal AL)

apk add nfs-utils

dmesg empy sofar

mount /home

dmesg

[4153944.457610] RPC: Registered named UNIX socket transport module.
[4153944.457615] RPC: Registered udp transport module.
[4153944.457618] RPC: Registered tcp transport module.
[4153944.457620] RPC: Registered tcp NFSv4.1 backchannel transport module.
[4153944.504475] svc: failed to register lockdv1 RPC service (errno 111).
[4153944.504484] lockd_up: makesock failed, error=-111
[4153944.681725] NFS: Registering the id_resolver key type
[4153944.681744] Key type id_resolver registered
[4153944.681748] Key type id_legacy registered

ls -ld /home

drwxr-xr-x    2 42949672 42949672        17 Jan 23 14:39 /home

cat /home/test

Do you see me?

touch /home/toto

touch: /home/toto: Permission denied

nfs-client in LXC AL (in KVM AL)

apk add nfs-utils

but

# mount /home
mount.nfs: rpc.statd is not running but is required for remote locking.
mount.nfs: Either use '-o nolock' to keep locks local, or start statd.
mount.nfs: an incorrect mount option was specified
mount: permission denied (are you root?)

and

# /etc/init.d/rpc.statd start
 * Caching service dependencies ... [ ok ]
 * Starting rpcbind ... [ ok ]
 * Starting NFS statd ... * start-stop-daemon: failed to start `/usr/sbin/rpc.statd'
 [ !! ]
 * ERROR: rpc.statd failed to start

dmesg

[74747.135827] rpcbind[6718]: segfault at 7ccfe7b0 ip 000072977ccef5cd sp 00007c6b3e329a68 error 4 in ld-musl-x86_64.so.1[72977cca0000+85000]
[74747.135841] grsec: Segmentation fault occurred at 000000007ccfe7b0 in /sbin/rpcbind[rpcbind:6718] uid/euid:100/100 gid/egid:101/101, parent /bin/busybox[init:1831] uid/euid:0/0 gid/egid:0/0
[74747.135887] grsec: bruteforce prevention initiated due to crash of /sbin/rpcbind against uid 100, banning suid/sgid execs for 15 minutes.  Please investigate the crash report for /sbin/rpcbind[rpcbind:6718] uid/euid:100/100 gid/egid:101/101, parent /bin/busybox[init:1831] uid/euid:0/0 gid/egid:0/0

nfs-client in LXC debian (in KVM AL)

apt-get install nfs-commonn

gives

[FAIL] Starting NFS common utilities: statd idmapd failed!

then mount /home gives same results in guest as in host

NFS-server in KVM-Alpine

Done from a KVM running in memory straight from the iso

CDROM="/my/path/alpine-mini-3.1.1-x86_64.iso"
qemu-system-x86_64 -name test -enable-kvm -cpu qemu64 -m 256 -smp 1 -curses \
  -net nic,vlan=0,model=virtio,macaddr=52:54:32:a0:a0:a0 \
  -net tap,vlan=0,script=/etc/openvswitch/ovs-ifup-lan,downscript=/etc/openvswitch/ovs-ifdown-lan,ifname=test0 \
  -cdrom ${CDROM}

do not forget to issue "grsec nomedeset" at SYSLINUX prompt or you loose the output (I'm doing it trough ssh term)

# setup-alpine # no disk install at all, no apk cache but proxy
# . /etc/profile.d/proxy.sh
# apk add nfs-utils
# echo "/home   192.168.1.0/24(rw,no_root_squash)" >> /etc/exports
# echo "Do you see me?" > /home/test
# /etc/init.d/nfs start
 * Caching service dependencies ...                                       [ ok ]
 * Starting rpcbind ...                                                   [ ok ]
 * Starting NFS statd ...
 * start-stop-daemon: failed to start `/usr/sbin/rpc.statd'               [ !! ]
 * ERROR: rpc.statd failed to start
 * ERROR: cannot start nfs as rpc.statd would not start
# dmesg # only relevant lines displayed
[  462.262020] rpcbind[1890]: segfault at 1e783940 ip 000070591e773f1d sp 00007dc1da01a4d8 error 4 in ld-musl-x86_64.so.1[70591e724000+86000]
[  462.262032] grsec: Segmentation fault occurred at 000000001e783940 in /sbin/rpcbind[rpcbind:1890] uid/euid:100/100 gid/egid:101/101, parent /bin/busybox[init:1] uid/euid:0/0 gid/egid:0/0              
[  462.262043] grsec: bruteforce prevention initiated due to crash of /sbin/rpcbind against uid 100, banning suid/sgid execs for 15 minutes.  Please investigate the crash report for /sbin/rpcbind[rpcbind:1890] uid/euid:100/100 gid/egid:101/101, parent /bin/busybox[init:1] uid/euid:0/0 gid/egid:0/0
# poweroff

Let's try with the vanilla kernel

CDROM="/my/path/alpine-vanilla-3.1.1-x86_64.iso"

with same command line and same sequence of instructions

test:~# /etc/init.d/nfs start
 * Caching service dependencies ...                                       [ ok ]
 * Starting rpcbind ...                                                   [ ok ]
 * Starting NFS statd ...
 * start-stop-daemon: failed to start `/usr/sbin/rpc.statd'               [ !! ]
 * ERROR: rpc.statd failed to start
 * ERROR: cannot start nfs as rpc.statd would not start
test:~# dmesg
[  243.445710] rpcbind[1930]: segfault at 33f30940 ip 00007f5a33f20f1d sp 00007fffa4290e48 error 4 in ld-musl-x86_64.so.1[7f5a33ed1000+86000] 
test:~# poweroff

Obviously I will not be able to test clients now...

nfs-client on bare metal AL

nfs-client in KVM AL

nfs-client in KVM debian

nfs-client in LXC AL (on bare metal AL)

nfs-client in LXC AL (in KVM AL)

nfs-client in LXC debian (in KVM AL)

How to automate KVM creation

The goal is not only to have a working install but to have it at the after setup-alpine stage without human intervention... Tis is the first stages of a work in progress...

I want to pass a Block Device and a name as parameters. The block device could be an image file, a LV, a NBD, a hdd, a raid array, whatever.
Everything else should be fully automatic according to some config file (stating the http-proxy, the time server, the log server, ...).

The I will just run the script, watch my dhcp logs to discover the new IP assigned (that's why the name is a parameter), then log in with ssh without password to customize it further but at high level only (will be a robot and not me in fact).

I guess it would be something like emulate boot from usb key with specific overlay already on key...
then run setup-disk with proper parameters on the command line to avoid the interactive process (like setup-alpine does)...
Methink this could be done from a couple of scripts put in /etc/local.d/. The last.stop one deleting all of them to be clean at next reboot.
Let's start easy ;)

How to prepare a img file to emulate an USB key

first a working example done in console (accessed trough ssh).
Will build a script from it...

First, lets's prepare somme block device (here an image file but could be something else)

apk add qemu-img
qemu-img create -f raw usbkey.img 512M
apk del qemu-img
T="usbkey.img"

Next, let's install AL on this $T

apk add multipath-tools syslinux dosfstools
fdisk $T
kpartx -av $T
mkdosfs -F32 /dev/mapper/loop1p1
dd if=/usr/share/syslinux/mbr.bin of=/dev/mapper/loop1
syslinux /dev/mapper/loop1p1
mkdir key
mount -t vfat /dev/mapper/loop1p1 key
wget http://wiki.alpinelinux.org/cgi-bin/dl.cgi/v3.1/releases/x86_64/alpine-mini-3.1.1-x86_64.iso
mkdir cdrom
mount alpine-mini-3.1.1-x86_64.iso cdrom
cd cdrom
cp -a .alpine-release * ../key/
cd ..
umount key
umount cdrom
kpartx -d $T
apk del multipath-tools syslinux dosfstools
rm alpine-mini-3.1.1-x86_64.iso