more pilot error

[William A. Mahaffey III]

I am having issues getting NFS server & client to work on my NetBSD
6.1.5 server: I see the following in the end of my messages file:


Jul 15 18:53:21 4256EE1 /netbsd: boot device: raid0
Jul 15 18:53:21 4256EE1 /netbsd: root on raid0a dumps on raid0b
Jul 15 18:53:21 4256EE1 /netbsd: root file system type: ffs
Jul 15 18:53:21 4256EE1 /netbsd: wsdisplay0: screen 1 added (80x25,
vt100 emulation)
Jul 15 18:53:21 4256EE1 /netbsd: wsdisplay0: screen 2 added (80x25,
vt100 emulation)
Jul 15 18:53:21 4256EE1 /netbsd: wsdisplay0: screen 3 added (80x25,
vt100 emulation)
Jul 15 18:53:21 4256EE1 /netbsd: wsdisplay0: screen 4 added (80x25,
vt100 emulation)
Jul 15 18:53:22 4256EE1 root: /etc/rc: WARNING: $rpcbind is not enabled.
Jul 15 18:53:22 4256EE1 root: /etc/rc: WARNING: /etc/exports is not
readable.
Jul 15 18:53:22 4256EE1 root: /etc/rc: WARNING: $rpcbind is not enabled.
Jul 15 18:53:22 4256EE1 savecore: /dev/rraid0b: Device not configured
Jul 15 18:53:22 4256EE1 timed[317]: slave to rpi
Jul 15 10:57:37 4256EE1 timed[317]: date changed by rpi from Wed Jul 15
18:53:22 2015
Jul 15 11:00:01 4256EE1 syslogd[159]: restart

2 (unrelated) issues, but I would like to get both resolved. I have no
dump device configured. I have 6 swap partitions:

4256EE1 # pstat -hs
Device Size Used Avail Capacity Priority
/dev/wd0e 16G 0B 16G 0% 0
/dev/wd1e 16G 0B 16G 0% 0
/dev/wd2e 16G 0B 16G 0% 0
/dev/wd3e 16G 0B 16G 0% 0
/dev/wd4e 16G 0B 16G 0% 0
/dev/wd5e 16G 0B 16G 0% 0
Total 96G 0B 96G 0%
4256EE1 # cat fstab
# NetBSD /targetroot/etc/fstab
# See /usr/share/examples/fstab/ for more examples.
/dev/raid0a / ffs rw 1 1
/dev/raid1a /usr ffs rw 1 2
/dev/wd0e none swap sw 0 0
/dev/wd1e none swap sw 0 0
/dev/wd2e none swap sw 0 0
/dev/wd3e none swap sw 0 0
/dev/wd4e none swap sw 0 0
/dev/wd5e none swap sw 0 0
/dev/dk0 /home ffs rw,log 1 3
kernfs /kern kernfs rw
ptyfs /dev/pts ptyfs rw
procfs /proc procfs rw
/dev/cd0a /cdrom cd9660 ro,noauto
/dev/sd0a /media/sd ufs rw,noauto
4256EE1 #

Is there a mount flag I can use to specify a dump device (preferably 1
of the swap partitions) ?

Also, I presume nfsd *can* work w/o rpcbind running, but I can't seem to
locate how to do that in the man pages. Any clues there appreciated. I
am on a LAN, this is *not* a public server, so only mount requests from
other LAN boxen will be processed. I (think I) configured everything as
needed in my rc.conf file:

4256EE1 # cat rc.conf
# $NetBSD: rc.conf,v 1.96 2000/10/14 17:01:29 wiz Exp $
#
# see rc.conf(5) for more information.
#
# Use program=YES to enable program, NO to disable it. program_flags are
# passed to the program on the command line.
#

# Load the defaults in from /etc/defaults/rc.conf (if it's readable).
# These can be overridden below.
#
if [ -r /etc/defaults/rc.conf ]; then
. /etc/defaults/rc.conf
fi

# If this is not set to YES, the system will drop into single-user mode.
#
rc_configured=YES

# Add local overrides below
#

hostname=4256EE1.CFD.COM
defaultroute="192.168.0.254"

sshd=YES
timed=YES
wscons=YES

# NFS daemons and parameters.
#
mountd=YES mountd_flags="" # NFS mount requests daemon
nfs_client=YES # enable client daemons
nfs_server=YES # enable server daemons
lockd=YES lockd_flags=""
statd=YES statd_flags=""

amd=YES
4256EE1 #

but the service doesn't start. More pilot error, most assuredly, but any
clues appreciated. TIA & have a good one.

--

William A. Mahaffey III

----------------------------------------------------------------------

"The M1 Garand is without doubt the finest implement of war
ever devised by man."
-- Gen. George S. Patton Jr.


--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-...@muc.de

[Greg Troxel]

rpcbind is used to translate service names to ports. Why do you not
want to run it? Before anything else, turn it on and reboot and only
when it works with try without.

[Brett Lymn]

On Wed, Jul 15, 2015 at 04:32:45PM -0453, William A. Mahaffey III wrote:
>
> dump device configured. I have 6 swap partitions:
>

That seems a bit excessive - the days of "2 * physical RAM plus a bit"
for swap allocation has not been needed for quite some time. I guess in
the general scheme of things the amount of storage you have allocated
for swap is not much but, still, it really doesn't need to be that much
IMHO. Also, why are they not a raid? At the
moment if you lose a disk that hosts the swap chances are the machine
will go down.

--
Brett Lymn

[William A. Mahaffey III]

Very well, I enabled it in my rc.conf file & did a 'rc.d/nfsd start' & I
now get the following line in my messages file:

Jul 15 19:30:12 4256EE1 nfsd[981]: can't register with udp portmap

The man page is mum about this, as (mostly) is Google. Any help appreciated.

[William A. Mahaffey III]

On 07/15/15 19:39, Brett Lymn wrote:
> On Wed, Jul 15, 2015 at 04:32:45PM -0453, William A. Mahaffey III wrote:
>> dump device configured. I have 6 swap partitions:
>>
> That seems a bit excessive - the days of "2 * physical RAM plus a bit"
> for swap allocation has not been needed for quite some time. I guess in
> the general scheme of things the amount of storage you have allocated
> for swap is not much but, still, it really doesn't need to be that much
> IMHO. Also, why are they not a raid? At the
> moment if you lose a disk that hosts the swap chances are the machine
> will go down.

Good questions. I have always heard that you get the best swasp
performance w/ the raw kernel driver (under SGI & linux), rather than
swapping to a RAID device. I have another box (Linux, FC14 64-bit) that
swaps to a RAID device, & it takes several min. to page in a few hundred
MB of paged-out-VM. That may be a linux only issue, but it is consistent
w/ my historical recollections. I have 96 GB of swap & 64 GB of RAM.
When fully operational the box may spend some time doing large CFD
calculations which exhaust that amount of RAM (pretty easy to do,
actually). The process should be able to continue, albeit more slowly,
which I will need. The swap devices are 1 slice on each of 6 HDD's, for
max parallelism & hopefully max performance. As soon as this machine is
fully up & running, I am going to redo the linux box to get rid of
swapping on a raid device (& allocate disk-space more usefully), & will
be going back to swapping on raw partitions due to crappy performance of
swap-on-raid on that box. The HDD's are brand new 2.5" drives, which I
have had good luck with historically, so I am hoping for more of the
same WRT HDD reliability vis-a-vis swapping (& other data partitions).

--

William A. Mahaffey III

----------------------------------------------------------------------

"The M1 Garand is without doubt the finest implement of war
ever devised by man."
-- Gen. George S. Patton Jr.

[Brett Lymn]

On Wed, Jul 15, 2015 at 07:48:41PM -0453, William A. Mahaffey III wrote:
> On 07/15/15 17:00, Greg Troxel wrote:
> >rpcbind is used to translate service names to ports. Why do you not
> >want to run it? Before anything else, turn it on and reboot and only
> >when it works with try without.
>
> Very well, I enabled it in my rc.conf file & did a 'rc.d/nfsd start' & I
> now get the following line in my messages file:
>
> Jul 15 19:30:12 4256EE1 nfsd[981]: can't register with udp portmap
>
> The man page is mum about this, as (mostly) is Google. Any help appreciated.
>

This is what I have in my rc.conf, nfs works fine:

# NFS daemons and parameters.
#

rpcbind=YES

mountd=YES mountd_flags="" # NFS mount requests
daemon
nfs_client=YES # enable client daemons
nfs_server=YES # enable server daemons

nfsd_flags="-6tun 4"

you probably don't need the client daemons. If it still doesn't work,
post up an output from rpcinfo.

--
Brett Lymn

[William A. Mahaffey III]

On 07/15/15 19:57, Brett Lymn wrote:
> On Wed, Jul 15, 2015 at 07:48:41PM -0453, William A. Mahaffey III wrote:
>> On 07/15/15 17:00, Greg Troxel wrote:
>>> rpcbind is used to translate service names to ports. Why do you not
>>> want to run it? Before anything else, turn it on and reboot and only
>>> when it works with try without.
>> Very well, I enabled it in my rc.conf file & did a 'rc.d/nfsd start' & I
>> now get the following line in my messages file:
>>
>> Jul 15 19:30:12 4256EE1 nfsd[981]: can't register with udp portmap
>>
>> The man page is mum about this, as (mostly) is Google. Any help appreciated.
>>
> This is what I have in my rc.conf, nfs works fine:
>
> # NFS daemons and parameters.
> #
> rpcbind=YES
> mountd=YES mountd_flags="" # NFS mount requests
> daemon
> nfs_client=YES # enable client daemons
> nfs_server=YES # enable server daemons
> nfsd_flags="-6tun 4"
>
> you probably don't need the client daemons. If it still doesn't work,
> post up an output from rpcinfo.
>

OK, here goes:

from my rc.conf:

# If this is not set to YES, the system will drop into single-user mode.
#
rc_configured=YES

# Add local overrides below
#

hostname=4256EE1.CFD.COM
defaultroute="192.168.0.254"

sshd=YES
timed=YES
wscons=YES

# NFS daemons and parameters.
#
rpcbind=YES
mountd=YES mountd_flags="" # NFS mount requests daemon

lockd=YES lockd_flags=""
statd=YES statd_flags=""

nfs_client=YES # enable client daemons

nfs_server=YES nfsd_flags="-6tun 4" # enable server daemons

amd=YES

& the results:

4256EE1 # rc.d/nfsd start
Starting nfsd.
4256EE1 # rc.d/nfsd status
nfsd is not running.
4256EE1 # rpcinfo
rpcinfo: Can't contact rpcbind (: RPC: Remote system error - No such
file or directory)
4256EE1 # uname -a
NetBSD 4256EE1.CFD.COM 6.1.5 NetBSD 6.1.5 (GENERIC) amd64
4256EE1 #

Same thing w/ the lockd & statd lines commented out. Thanks & TIA ....

--

William A. Mahaffey III

----------------------------------------------------------------------

"The M1 Garand is without doubt the finest implement of war
ever devised by man."
-- Gen. George S. Patton Jr.

[Brett Lymn]

On Thu, Jul 16, 2015 at 12:02:43AM -0453, William A. Mahaffey III wrote:
>
> 4256EE1 # rc.d/nfsd start
> Starting nfsd.
> 4256EE1 # rc.d/nfsd status
> nfsd is not running.
> 4256EE1 # rpcinfo
> rpcinfo: Can't contact rpcbind (: RPC: Remote system error - No such
> file or directory)
>

OK, this is bad. Is rpcbind actually running? You have to have that
for nfs to work. Work on getting rpcinfo returning valid information,
that is the foundation for everything else. Once rpcinfo is reporting
something like this:

[blymn@siren] rpcinfo
program version netid address service owner
100000 4 tcp 0.0.0.0.0.111 portmapper superuser
100000 3 tcp 0.0.0.0.0.111 portmapper superuser
100000 2 tcp 0.0.0.0.0.111 portmapper superuser
100000 4 udp 0.0.0.0.0.111 portmapper superuser
100000 3 udp 0.0.0.0.0.111 portmapper superuser
100000 2 udp 0.0.0.0.0.111 portmapper superuser
100000 4 tcp6 ::.0.111 portmapper superuser
100000 3 tcp6 ::.0.111 portmapper superuser
100000 4 udp6 ::.0.111 portmapper superuser
100000 3 udp6 ::.0.111 portmapper superuser
100000 4 local /var/run/rpcbind.sock portmapper superuser
100000 3 local /var/run/rpcbind.sock portmapper superuser
100000 2 local /var/run/rpcbind.sock portmapper superuser
100005 1 udp 0.0.0.0.3.254 mountd superuser
100005 3 udp 0.0.0.0.3.254 mountd superuser
100005 1 tcp 0.0.0.0.3.255 mountd superuser
100005 3 tcp 0.0.0.0.3.255 mountd superuser
100005 1 udp6 ::.3.254 mountd superuser
100005 3 udp6 ::.3.254 mountd superuser
100005 1 tcp6 ::.3.255 mountd superuser
100005 3 tcp6 ::.3.255 mountd superuser
100003 2 udp 0.0.0.0.8.1 nfs superuser
100003 3 udp 0.0.0.0.8.1 nfs superuser
100003 2 tcp 0.0.0.0.8.1 nfs superuser
100003 3 tcp 0.0.0.0.8.1 nfs superuser
100003 2 udp6 ::.8.1 nfs superuser
100003 3 udp6 ::.8.1 nfs superuser
100003 2 tcp6 ::.8.1 nfs superuser
100003 3 tcp6 ::.8.1 nfs superuser

then it is likely that nfs will Just Work (tm).

[William A. Mahaffey III]

Hmmmmm .... OK, I manually started rpcbind & it started up fine, then
nfsd & it started up fine. I am new to NetBSD, didn't know I had to
separately, explicitly start rpcbind, I thought the nfsd start process
would do that. Live & learn. Did I mention I am a *noob* to NetBSD :-) ?
Thanks, I seem to be past this issue.

--

William A. Mahaffey III

----------------------------------------------------------------------

"The M1 Garand is without doubt the finest implement of war
ever devised by man."
-- Gen. George S. Patton Jr.

[Brett Lymn]

On Thu, Jul 16, 2015 at 08:08:20AM -0453, William A. Mahaffey III wrote:
>
> Hmmmmm .... OK, I manually started rpcbind & it started up fine, then
> nfsd & it started up fine. I am new to NetBSD, didn't know I had to
> separately, explicitly start rpcbind, I thought the nfsd start process
> would do that. Live & learn. Did I mention I am a *noob* to NetBSD :-) ?
> Thanks, I seem to be past this issue.
>

Everyone starts off somewhere. Your assumption is not unreasonable, it
doesn't make a whole lot of sense to have nfsd start without rpcbind
running. We probably should ensure rpcbind is running in the nfsd rc
script. I can see there is a dependency there but I don't think
anything actually ensures it is running. I guess because the assumption
is that the dependencies on boot would have started it beforehand.

Anyway, glad you are over that hurdle.

[William A. Mahaffey III]

On 07/16/15 19:56, Brett Lymn wrote:

> On Thu, Jul 16, 2015 at 08:08:20AM -0453, William A. Mahaffey III wrote:
>> Hmmmmm .... OK, I manually started rpcbind & it started up fine, then
>> nfsd & it started up fine. I am new to NetBSD, didn't know I had to
>> separately, explicitly start rpcbind, I thought the nfsd start process
>> would do that. Live & learn. Did I mention I am a *noob* to NetBSD :-) ?
>> Thanks, I seem to be past this issue.
>>

> Everyone starts off somewhere. Your assumption is not unreasonable, it
> doesn't make a whole lot of sense to have nfsd start without rpcbind
> running. We probably should ensure rpcbind is running in the nfsd rc
> script. I can see there is a dependency there but I don't think
> anything actually ensures it is running. I guess because the assumption
> is that the dependencies on boot would have started it beforehand.
>
> Anyway, glad you are over that hurdle.

Yep, absolutely correct, everything would be AOK on boot, but it is nice
to be able to restart services on the fly relatively simply, which is
what I was expecting. Like I said, live & learn. No problema :-) ....

--

William A. Mahaffey III

----------------------------------------------------------------------

"The M1 Garand is without doubt the finest implement of war
ever devised by man."
-- Gen. George S. Patton Jr.

[Jan Danielsson]

On 17/07/15 02:49, Brett Lymn wrote:
>> Hmmmmm .... OK, I manually started rpcbind & it started up fine, then
>> nfsd & it started up fine. I am new to NetBSD, didn't know I had to
>> separately, explicitly start rpcbind, I thought the nfsd start process
>> would do that. Live & learn. Did I mention I am a *noob* to NetBSD :-) ?
>> Thanks, I seem to be past this issue.
>

> Everyone starts off somewhere. Your assumption is not unreasonable, it
> doesn't make a whole lot of sense to have nfsd start without rpcbind
> running. We probably should ensure rpcbind is running in the nfsd rc
> script. I can see there is a dependency there but I don't think
> anything actually ensures it is running. I guess because the assumption
> is that the dependencies on boot would have started it beforehand.

Yes, that's it. Those dependency rules are there to help rcorder(8)
get the order right during startup, but when you start an rc.d script
manually that interdependency metadata is no more than comments in a script.

--
Kind Regards,
Jan