nfs related warnings in log files
knireis
For the syslog:
Dec 11 21:36:10 DNS323 user.warn kernel: svc: 192.168.2.100, port=768:
unknown version (4 for prog 100003, nfsd)
Dec 11 21:36:10 DNS323 user.warn kernel: svc: 192.168.2.100, port=832:
unknown version (4 for prog 100003, nfsd)
Dec 11 21:36:10 DNS323 user.warn kernel: svc: 192.168.2.100, port=843:
unknown version (4 for prog 100003, nfsd)
Dec 11 21:36:10 DNS323 user.warn kernel: svc: 192.168.2.100, port=705:
unknown version (4 for prog 100003, nfsd)
Dec 11 21:36:10 DNS323 user.warn kernel: svc: 192.168.2.100, port=809:
unknown version (4 for prog 100003, nfsd)
Dec 11 21:36:10 DNS323 user.warn kernel: svc: 192.168.2.100, port=769:
unknown version (4 for prog 100003, nfsd)
Dec 11 21:36:10 DNS323 user.warn kernel: svc: 192.168.2.100, port=940:
unknown version (4 for prog 100003, nfsd)
Dec 11 21:36:10 DNS323 user.warn kernel: svc: 192.168.2.100, port=824:
unknown version (4 for prog 100003, nfsd)
Dec 11 21:36:10 DNS323 user.warn kernel: svc: 192.168.2.100, port=672:
unknown version (4 for prog 100003, nfsd)
Dec 11 21:36:10 DNS323 user.warn kernel: svc: 192.168.2.100, port=921:
unknown version (4 for prog 100003, nfsd)
and for the kernel log:
svc: 192.168.2.100, port=966: unknown version (4 for prog 100003,
nfsd)
svc: 192.168.2.100, port=873: unknown version (4 for prog 100003,
nfsd)
svc: 192.168.2.100, port=869: unknown version (4 for prog 100003,
nfsd)
svc: 192.168.2.100, port=976: unknown version (4 for prog 100003,
nfsd)
svc: 192.168.2.100, port=926: unknown version (4 for prog 100003,
nfsd)
svc: 192.168.2.100, port=775: unknown version (4 for prog 100003,
nfsd)
svc: 192.168.2.100, port=676: unknown version (4 for prog 100003,
nfsd)
svc: 192.168.2.100, port=686: unknown version (4 for prog 100003,
nfsd)
svc: 192.168.2.100, port=922: unknown version (4 for prog 100003,
nfsd)
I use auto fs from my laptop (linux mint, ubuntu, 192.168.2.100) to
access the files on the nas. I suspect these warnings to prevent my
disk to go to sleep. But when the disk is in sleep, it does not wakeup
because of these warnings.
Joao Cardoso
On Dec 12, 3:55 pm, knireis <mail...@jsierink.nl> wrote:
> My log files are filled up with these kind of warnings:
> For the syslog:
> Dec 11 21:36:10 DNS323 user.warn kernel: svc: 192.168.2.100, port=768:
> unknown version (4 for prog 100003, nfsd)
> svc: 192.168.2.100, port=966: unknown version (4 for prog 100003,
> nfsd)
If you use on you laptop "rpcinfo <your_dns>", you will see that all
Alt-F nfs components (but lockmanager) advertise version 3, so why is
the laptop requesting a non-advertised version?!
> I use auto fs from my laptop (linux mint, ubuntu, 192.168.2.100) to
> access the files on the nas.
nfs v4 does not fit the available flash memory space, as additional
kernel modules and two more libraries are needed. Perhaps after
removing libncurses and emacs? Are there other candidates to remove/
add from the firmware?
> I suspect these warnings to prevent my
> disk to go to sleep.
up), so that is not the reason.
But prevents? It does not go to sleep when no other PC is alive at
home? After the timeout settings in disk->utilities? (don't forget to
"Submit")
>But when the disk is in sleep, it does not wakeup
> because of these warnings.
the ntp server? cron?... is swap being used?
knireis
On 12 dec, 18:27, Joao Cardoso <whoami.jc...@gmail.com> wrote:
> On Dec 12, 3:55 pm, knireis <mail...@jsierink.nl> wrote:
>
> > My log files are filled up with these kind of warnings:
> > For the syslog:
> > Dec 11 21:36:10 DNS323 user.warn kernel: svc: 192.168.2.100, port=768:
> > unknown version (4 for prog 100003, nfsd)
> ...
> > and for the kernel log:
> > svc: 192.168.2.100, port=966: unknown version (4 for prog 100003,
> > nfsd)
>
> Alt-F only supports NFS v3, and your laptop is trying to use v4.
>
> If you use on you laptop "rpcinfo <your_dns>", you will see that all
> Alt-F nfs components (but lockmanager) advertise version 3, so why is
> the laptop requesting a non-advertised version?!
>
> > I use auto fs from my laptop (linux mint, ubuntu, 192.168.2.100) to
> > access the files on the nas.
>
> Try using vers=3 as one of the automount nfs options
flooded with messages like:
Dec 13 10:13:06 DNS323 daemon.info portmap[1460]: connect from
192.168.2.3 to getport(100005)
Dec 13 10:13:06 DNS323 daemon.info portmap[1459]: connect from
192.168.2.3 to getport(100003)
Dec 13 10:13:06 DNS323 daemon.info portmap[1461]: connect from
192.168.2.3 to getport(100003)
Dec 13 10:13:06 DNS323 daemon.notice rpc.mountd[815]: authenticated
mount request from 192.168.2.3:971 for /mnt/aaa (/mnt/aaa)
Dec 13 10:13:13 DNS323 daemon.info portmap[1469]: connect from
192.168.2.3 to getport(100005)
So i switched to using smb and mount the nas via the autofs package on
my laptop, This does not create entries in the log.
>
> > I suspect these warnings to prevent my
> > disk to go to sleep.
>
> Logs use 32KB of memory (what a waste!), not the disk (to not spin it
> up), so that is not the reason.
spins up with no apparent reason (no computer in the network is on)
> But prevents? It does not go to sleep when no other PC is alive at
> home? After the timeout settings in disk->utilities? (don't forget to
> "Submit")
>
> >But when the disk is in sleep, it does not wakeup
> > because of these warnings.
>
> Perhaps other process is accesing the disk? The web status page. e.g?
> the ntp server? cron?... is swap being used?
some time (a lot more then the time specified in the webinterface) I
will make a separate topic about spinup/down
Joao Cardoso
> On 12 dec, 18:27, Joao Cardoso <whoami.jc...@gmail.com> wrote:
> This solution works. No more warnings in the log, but the log is now
> flooded with messages like:
> Dec 13 10:13:06 DNS323 daemon.info portmap[1460]: connect from
> 192.168.2.3 to getport(100005)
> Dec 13 10:13:06 DNS323 daemon.info portmap[1459]: connect from
> 192.168.2.3 to getport(100003)
> Dec 13 10:13:06 DNS323 daemon.info portmap[1461]: connect from
> 192.168.2.3 to getport(100003)
> Dec 13 10:13:06 DNS323 daemon.notice rpc.mountd[815]: authenticated
> mount request from 192.168.2.3:971 for /mnt/aaa (/mnt/aaa)
> Dec 13 10:13:13 DNS323 daemon.info portmap[1469]: connect from
> 192.168.2.3 to getport(100005)
Unfortunately B5 was shipped with portmap in verbose mode.
Those messages are harmless (as the others also was) -- that is only portmap
reporting that someone established communication with him to contact the
several nfs server components.
You see, the local portmap accepts registration from the local nfs server
componens, and then advertises it in the network (that is what you saw in the
host client with the rpcinfo command). When the client wants to establish
contact with a specific server component, it contacts again portmap asking for
a dinamically assigned port number -- that is what you are seeing now
Automount on your laptop has to contact the portmap (and nfs server
components, but those are not in verbose mode) to mount the "share". But
automount unmount the share after a certain period of inactivity, and than
mounts it again, that is why you see so many messages in the log.
Whatch the log times carefully
As the logs are in memory and rotate (max of 32KB), I expect performance to
not be degraded.
If your box is connected 24/7, you can to turnoff portmap verbosity. Just edit
/etc/init.d/S60nfs
and comment the line who says
PORTMAP_FLAG=-v
This will not survive a reboot, of course, unless you are willing to read and
understand the
HowToFixOrCustomizeFirmware wiki entry.
> So i switched to using smb and mount the nas via the autofs package on
> my laptop, This does not create entries in the log.
>
> > > I suspect these warnings to prevent my
> > > disk to go to sleep.
...
> By using smb in stead of nfs my disk spins up less often. But it still
> spins up with no apparent reason (no computer in the network is on)
that's tipical of linux. Some measures can be taken to alliviate the problem,
but they are not 100% effective.
nfs has another characteristic that might contribute to the problem: the
server stores some status on disk, so that after a crash/porweroff/poweron it
can resume operations and continue serving the same clients, without they even
noticed what happened (but the clients blocks if they try to contact the
server -- when it comes back the client de-blocks and continues).
With this nfs philosofy in mind, Alt-F sets the save-status directory on disk.
This was a design decision, as I could setup those directories in memory,
trying to avoid the disk spinup. The problem is that on a server crash you
would have to poweroff the clients also, as the server could not resume
operations.
Unless... (this is becaming a class!) that in the client you mounted the nfs
share in "soft" mode (not "hard" -- but then you could have data loss...
Joao