View Full Version : [ET10x00] box invisible on network
Having an issue with box not seen on network even with nfs and samba running,rebooted and powered off/on.
Every box setup mounts are done when seen but for some reason it does not find it under network scan:confused:
With this new image and now password needed could that be blocking it;)
I dont use passwords when mounting hdds.
Just to add other boxes also dont see the et.
Have you gone into the ET settings and ran a network test?
You don't PW for network discovery but will do to share HDD and other features via remote access
Sent from my Mi MIX 3 5G using Tapatalk
Yes all in the green as just rechecked and no logs as of yet to show anything.
BrokenUnusableAccount
25-12-20, 14:55
I've always done it with just SAMBA and no NFS (NFS not installed).
If you're trying to browse the box from Windows see if you can do it just by IP address, put something like:
\\192.168.1.42\Harddisk\
in to the address bar at the top of an explorer window and press enter. (but use the IP address of your box on the LAN)
If it asks for a password I usually give it root and my password but, as I understand it, you probably don't need to if you only want read access.
same IP range ? Same workgroup ?
No pc or laptop used just e2 boxes to access hdd on et10000.
Even the et10000 itself cannot find itself with a network browser scan which it did before on older image that went slow with errors.
Only thing has changed is the latest image is in use now and boxes still in same place as before.
abu baniaz
26-12-20, 01:06
I am pretty sure you will need to set passwords
I used a password on filezilla to gain access to box,but et10000 does not see itself on a network scan which is very odd,thinking its new image related?
On laptop i ran a webif control of a network browse and no et10000
abu baniaz
26-12-20, 03:23
Set password on box.
https://www.world-of-satellite.com/showthread.php?63400-OpenViX-5-4-released!&p=503909&viewfull=1#post503909
Password is on box and retried a rescan and still not seen and box has been rebooted as just checked.
I did find new logs as might be of use?
I had an issue last week where after a settings restore (which restored password) I couldn‘t access the box - changed password to same as before and everything ran OK. If you haven‘t reset password worth trying.
Password is on box and retried a rescan and still not seen and box has been rebooted as just checked.
I did find new logs as might be of use?Try setting password through telnet.
I had issue where box was seen on network but could not access it with the remote set on the box, telneted and setup password, FileZilla connected.
Think it would be strange for network discovery to be disabled if no password set but logic would make sense from a security perspective.
Sent from my Mi MIX 3 5G using Tapatalk
I dont access network from outside like some do,this was a fresh flash and setup from scratch and had to add a password to use filezilla.
I also changed it again as i can login to box and use remote so i think telnet might not be of use.
My box is invisible on network and only access to hdd is in bedroom as where box is situated.
Would network discovery be shown in logs?
My 2 boxes can see themselves using network browser, but both then insist on a password being set first (via the browser).
The password was already set on both in the browser, but needed saving again (has been mentioned before).
What is odd/wrong is that the browser and mount_manager/user_management had different usernames set, the browser (I think?) was still username (and worked!), the other was root.
I'm wondering if accessing the boxes from a pc first (using \\ip address) is needed to kick start network browsing. Just a thought.
But i can use webif to log into box and control it via laptop,and saved using green button:confused:
Hence why i think might be a bug?
But i can use webif to log into box and control it via laptop,and saved using green button:confused:
Hence why i think might be a bug?
That doesn't use Samba/Nfs
Does Windows_Explorer/Network see anything? If not type \\box_ip_address in the box to the left of search_network , and then try network_browser on ViX again.
Ok tried laptop search which lists all network connections and et10000 and others are listed,clicking on the zero can access box but clicking on et does not enter and says a network error.Went into network and restarted network connections and ran a new scan still not seen.
Network error code 0x80070035
the network path not found.
Typing in box ip/address/harddrive/ i get a windows error 404 in using browser.
Indulge me, telnet into the box, set the password and then retry connecting etc.
Sent from my Mi MIX 3 5G using Tapatalk
BrokenUnusableAccount
26-12-20, 15:34
Typing in box ip/address/harddrive/ i get a windows error 404 in using browser.
Well yes. That's never going to work in a web browser.
I wanted you to try:
\\192.168.1.42\Harddisk\
in a file explorer window.
Indulge me, telnet into the box, set the password and then retry connecting etc.
Sent from my Mi MIX 3 5G using Tapatalk
Unsure if doing it correctly but nothing happens and entered root and passwords i used before.
in a file explorer window.
Used the search parameters on lower left of screen to look for file
Open up CMD
Type telnet IP address of box
To change password type passed
You might need to enable telnet or download a telnet app
Sent from my Mi MIX 3 5G using Tapatalk
BrokenUnusableAccount
26-12-20, 17:08
type passed
That should be:
passwd
Password entered with passwd and saved and exited rebooted box and rescanned box is still invisible lol.
are you playing with us :) ? Quote : No pc or laptop used just e2 boxes to access hdd on et10000. Dispite even if samba or NFS is not active can you see the box in your router ? Think you have a router between boxes or simply a switch
PS ::maybe explain just how everything is connected and installed included ip adresses
are you playing with us :) ? Quote : No pc or laptop used just e2 boxes to access hdd on et10000. Dispite even if samba or NFS is not active can you see the box in your router ? Think you have a router between boxes or simply a switch
PS ::maybe explain just how everything is connected and installed included ip adresses
I have a pc and laptop but only used with a connection to box.I dont use pc or laptop to view recordings on them as just access them from other boxes like i did before.
Virgin router and unable to log into router until i find password upstairs -bad memory.
4 ports with cables going to multiple rooms with multiple switches.
Boxes all have lan cables and hard wired with DHCP used for the last 10 years and all have gateway and mac address.
If i search for recordings on this box from my zgemma from its mount Harddisk Et10000 it does not exist.
Never had an issue with an invisible box not showing up on its own search?
Hence why it might be a bug as i think all bases are covered.
Previous images had no issues with its footprint and box is on recording mode now.
maybe than go back to previous image for testing.
Yes might have to go back 2 years as i think all avenues have been checked and not using the last one that was corrupted.
Restored old image and its there in a network scan on et and also shows up now on my h7s
Does that confirm a bug?
Restored old image and its there in a network scan on et and also shows up now on my h7s
Does that confirm a bug?
Difficult to say, my ET10K works as expected using Samba. (Did you check that Samba was running, I've noticed it sometimes is in the "stopped" state for some unknown reason.)
My ET10k has always been updated (every release) by couch flash and a settings restore. The last build from scratch was when it was mandatory - 4. something.
Have you changed the Workgroup at some time, the default "WORKGROUP" ?
If i cant access a box i check to see if Samba has stopped which it has and i restart it again.No not changed 'Workgroup'
I only reflash when needed and box had an issue with previous image.
I only set boxes up with fresh images and no settings restore but in some cases will have to use restore as in old image which is on box now.
BrokenUnusableAccount
27-12-20, 02:59
I only set boxes up with fresh images and no settings restore but in some cases will have to use restore as in old image which is on box now.
I wish you'd tried restoring your settings.
That seems to be the only different thing that most people do but you don't do.
Maybe try again at some point with a later 5.4.xxx version.
I wish you'd tried restoring your settings.
That seems to be the only different thing that most people do but you don't do.
Maybe try again at some point with a later 5.4.xxx version.
I did and box still had issues from previous bug with vix spinner is why i tried it first and eventually had to do a clean fresh install.
I've reflashed my ET10K with 5.4.003, no settings restore, started Samba (needed an opkg update to get it to load ok.)
PC could see ET10K (asked for login credentials).
With browse network neighbourhood, ET10K could see Ultimo4k, but not itself. Ultimo4k could not see the ET10K.
The mounts I have on the Ultimo4k include the parameter vers=2.0 I couldn't work out how to force this onto browse network, so.......
Took a copy of /etc/samba from the working version of 5.4.003, and copied it into the non-working version. Now works ok (needed a yellow button rescan in browser.)
(To copy, I stopped enigma (telnet init 4), renamed /etc/samba to /etc/samba.orig, copied over the "old" version of /etc/samba, restarted with init 3, and then rebooted.)
So it must be a bug,now to try what you did Cheers
So it must be a bug,now to try what you did Cheers
I suspect the longer term solution is to mount the boxes manually, no big deal, and avoids trying to work round limitations which may never/can never be fixed.
But if the et is not seen it cant be mounted and box only sees media hdd,Might be fixed in next image lol.
When you said mount it manually via remote with box in front do you mean?
Use mount manager to tell it exactly what you want to do.....
Tried that and still unseen, hmm not sure if i entered a password as i think i did:confused:
Tried doing it on my h7s and box crashed as it sees the box but cannot enter the box.
Tried that and still unseen, hmm not sure if i entered a password as i think i did:confused:
Haven't you got network browser working again, that's the issue, it struggles talking to different samba versions.
Not done the init thing yet,ahh with you now samba needs an update/adjustment.
BrokenUnusableAccount
27-12-20, 21:10
I've reflashed my ET10K with 5.4.003, no settings restore, started Samba (needed an opkg update to get it to load ok.)
PC could see ET10K (asked for login credentials).
With browse network neighbourhood, ET10K could see Ultimo4k, but not itself. Ultimo4k could not see the ET10K.
The mounts I have on the Ultimo4k include the parameter vers=2.0 I couldn't work out how to force this onto browse network, so.......
Took a copy of /etc/samba from the working version of 5.4.003, and copied it into the non-working version. Now works ok (needed a yellow button rescan in browser.)
(To copy, I stopped enigma (telnet init 4), renamed /etc/samba to /etc/samba.orig, copied over the "old" version of /etc/samba, restarted with init 3, and then rebooted.)
Are you able to copy /etc/samba.orig and /etc/samba to a PC and compare the contents?
And/or ZIP them and upload so others can compare and experiment?
Are you able to copy /etc/samba.orig and /etc/samba to a PC and compare the contents?
And/or ZIP them and upload so others can compare and experiment?
I did a bit of checking after I'd finished, but found no differences, probably because I was rushing and mixed up the versions.
I'll have another look tomorrow.
I have not done anything as recording a load of things again or fell asleep:rolleyes:
I had an idea, unusual I know, more surprising - it worked.
Flashed ET10K with 5.4.003, no settings restore. Started Samba (after opkg update) and network browser exhibited the documented problems.
Did a settings restore and then restored settings+plugins - Samba now working as expected.:)
There is one subtle difference with the samba directory before and after the settings restore.......
root@et10000:~# ls -lR /etc/samba*
/etc/samba.flash.orig:
drwxr-xr-x 2 root root 528 Dec 28 09:02 distro
-rw-r--r-- 1 root root 20 Dec 28 09:02 lmhosts
drwxr-xr-x 2 root root 376 Dec 28 09:02 private
-rw-r--r-- 1 root root 744 Dec 28 09:02 smb-user.conf
-rw-r--r-- 1 root root 95 Dec 28 09:02 smb.conf
/etc/samba.flash.orig/distro:
-rw-r--r-- 1 root root 61 Dec 28 09:02 smb-branding.conf
-rw-r--r-- 1 root root 1457 Dec 28 09:02 smb-global.conf
-rw-r--r-- 1 root root 586 Dec 28 09:02 smb-shares.conf
-rw-r--r-- 1 root root 29 Dec 28 09:02 smb-vmc.conf
-rw-r--r-- 1 root root 29 Dec 28 09:02 smb-vmc.samba
/etc/samba.flash.orig/private:
-rw-r--r-- 1 root root 430080 Dec 28 09:02 secrets.tdb
-rw-r--r-- 1 root root 204 Dec 28 09:02 smbpasswd
-rw-r--r-- 1 root root 80 Dec 28 09:02 users.map
root@et10000:~#
after settings restore.....
root@et10000:~# ls -lR /etc/samba
/etc/samba:
drwxr-xr-x 2 root root 528 Dec 28 09:06 distro
-rw-r--r-- 1 root root 20 Oct 8 05:12 lmhosts
drwxr-xr-x 2 root root 376 Dec 28 09:06 private
-rw-r--r-- 1 root root 744 Oct 8 05:12 smb-user.conf
-rw-r--r-- 1 root root 95 Oct 8 05:12 smb.conf
/etc/samba/distro:
-rw-r--r-- 1 root root 61 Oct 8 05:12 smb-branding.conf
-rw-r--r-- 1 root root 1457 Oct 8 05:12 smb-global.conf
-rw-r--r-- 1 root root 586 Oct 8 05:12 smb-shares.conf
lrwxrwxrwx 1 root root 13 Dec 28 09:06 smb-vmc.conf -> smb-vmc.samba
-rw-r--r-- 1 root root 29 Oct 8 05:12 smb-vmc.samba
/etc/samba/private:
-rw------- 1 root root 430080 Dec 28 08:57 secrets.tdb
-rw------- 1 root root 204 Dec 28 08:57 smbpasswd
-rw-r--r-- 1 root root 80 Oct 8 05:12 users.map
root@et10000:~#
So restoring the plugin produces the desired result, no need to use an old working copy of /etc/samba (although the effect is the same).
Anything else needed before I get back to the full image?
My vhannibal has thrown a wobbly reset my sat tuners and not listed bouquets and and throwing up an erno warning .
Need to start again as cant remember with all the restores which one was used.
I had an idea, unusual I know, more surprising - it worked.
Flashed ET10K with 5.4.003, no settings restore. Started Samba (after opkg update) and network browser exhibited the documented problems.
Did a settings restore and then restored settings+plugins - Samba now working as expected.:)
There is one subtle difference with the samba directory before and after the settings restore.......
root@et10000:~# ls -lR /etc/samba*
/etc/samba.flash.orig:
drwxr-xr-x 2 root root 528 Dec 28 09:02 distro
-rw-r--r-- 1 root root 20 Dec 28 09:02 lmhosts
drwxr-xr-x 2 root root 376 Dec 28 09:02 private
-rw-r--r-- 1 root root 744 Dec 28 09:02 smb-user.conf
-rw-r--r-- 1 root root 95 Dec 28 09:02 smb.conf
/etc/samba.flash.orig/distro:
-rw-r--r-- 1 root root 61 Dec 28 09:02 smb-branding.conf
-rw-r--r-- 1 root root 1457 Dec 28 09:02 smb-global.conf
-rw-r--r-- 1 root root 586 Dec 28 09:02 smb-shares.conf
-rw-r--r-- 1 root root 29 Dec 28 09:02 smb-vmc.conf
-rw-r--r-- 1 root root 29 Dec 28 09:02 smb-vmc.samba
/etc/samba.flash.orig/private:
-rw-r--r-- 1 root root 430080 Dec 28 09:02 secrets.tdb
-rw-r--r-- 1 root root 204 Dec 28 09:02 smbpasswd
-rw-r--r-- 1 root root 80 Dec 28 09:02 users.map
root@et10000:~#
after settings restore.....
root@et10000:~# ls -lR /etc/samba
/etc/samba:
drwxr-xr-x 2 root root 528 Dec 28 09:06 distro
-rw-r--r-- 1 root root 20 Oct 8 05:12 lmhosts
drwxr-xr-x 2 root root 376 Dec 28 09:06 private
-rw-r--r-- 1 root root 744 Oct 8 05:12 smb-user.conf
-rw-r--r-- 1 root root 95 Oct 8 05:12 smb.conf
/etc/samba/distro:
-rw-r--r-- 1 root root 61 Oct 8 05:12 smb-branding.conf
-rw-r--r-- 1 root root 1457 Oct 8 05:12 smb-global.conf
-rw-r--r-- 1 root root 586 Oct 8 05:12 smb-shares.conf
lrwxrwxrwx 1 root root 13 Dec 28 09:06 smb-vmc.conf -> smb-vmc.samba
-rw-r--r-- 1 root root 29 Oct 8 05:12 smb-vmc.samba
/etc/samba/private:
-rw------- 1 root root 430080 Dec 28 08:57 secrets.tdb
-rw------- 1 root root 204 Dec 28 08:57 smbpasswd
-rw-r--r-- 1 root root 80 Oct 8 05:12 users.map
root@et10000:~#
So restoring the plugin produces the desired result, no need to use an old working copy of /etc/samba (although the effect is the same).
Anything else needed before I get back to the full image?
I tried
root@et10000:~# opkg list | grep -e smb -e samba before and after the settings restore, the results were identical.
I'm assuming all samba packages include the text smb or samba.
I'm giving up, I can no longer get the simple sequence in post #45 to work.
I started again from scratch with debug logs on, after that I couldn't get network browser to work as expected.
The exception file straight after the flash and samba install now shows as a symbolic link, not as it did in #45....
-rw-r--r-- 1 root root 29 Dec 28 09:02 smb-vmc.conf
There's nothing I could have done to make that happen.
Taking a settings backup and restoring it no longer did what it achieved in #45.
I restored settings+plugins from a working backup and network browser works ok.
BrokenUnusableAccount
28-12-20, 15:57
One final thought.
Could it be that nmbd stops running for some unknown reason?
This command should show the smbd and nmbd processes that are running.
ps -AFH | grep -e [sn]mbd
There should be an smbd (and it's child smbd processes) and an nmbd process.
On by Zgemma H7 I got this:
root@zgemmah7:~#
root@zgemmah7:~# ps -AFH | grep -e [sn]mbd
root 2061 1 0 7251 2992 1 Dec27 ? 00:00:00 /usr/sbin/smbd
root 2065 2061 0 6600 1960 1 Dec27 ? 00:00:00 /usr/sbin/smbd
root 2066 2061 0 6601 2228 0 Dec27 ? 00:00:00 /usr/sbin/smbd
root 2068 1 0 4355 1468 0 Dec27 ? 00:00:03 /usr/sbin/nmbd
root@zgemmah7:~#
As I understand it the important thing is that both smbd and nmbd are running.
The number of extra child smbd processes varies depending on, I presume, present (and past?) connections that have been made to samba.
I found during my last attempt(s) samba had sometimes stopped and I had to restart it via the gui.
This is my current "live" working version on the et10k...
root@et10000:~# ps -AFH | grep -e [sn]mbd
root 604 1 0 8012 7864 0 13:27 ? 00:00:00 /usr/sbin/smbd
root 608 604 0 7293 2876 0 13:27 ? 00:00:00 /usr/sbin/smbd
root 609 604 0 7291 2580 0 13:27 ? 00:00:00 /usr/sbin/smbd
root 611 1 0 4794 2952 0 13:27 ? 00:00:00 /usr/sbin/nmbd
root@et10000:~#
Maybe I'll have one last go, I must be mad.
@ccs thats why we love ya:fart2::thumbsup:
As I understand it the important thing is that both smbd and nmbd are running.
The number of extra child smbd processes varies depending on, I presume, present (and past?) connections that have been made to samba.
If I kill nmbd, I still see 3 other assorted boxes, but not the et10k.
Restart nmbd and the et10k comes back.
May i ask what are smbd abn nmbd please Cheers.
BrokenUnusableAccount
28-12-20, 17:30
May i ask what are smbd abn nmbd please Cheers.
They are parts of Samba.
https://www.samba.org/samba/docs/current/man-html/smbd.8.html
https://www.samba.org/samba/docs/current/man-html/nmbd.8.html
If I kill nmbd, I still see 3 other assorted boxes, but not the et10k.
Restart nmbd and the et10k comes back.
Back to square one, reflashed 5.4.003, no settings restore, start samba, no /usr/sbin/nmbd running.
Start /usr/sbin/nmbd by hand, network browser working fine.
So need to find out what autostarts /usr/sbin/nmbd (and now samba has been in the stopped state after 3 reboots, so probably part of the same rc file).
There's a startup script in /etc/rc0_to_6.d. I'm assuming they're all similar, named either K20samba or S20samba.
#! /bin/sh
### BEGIN INIT INFO
# Provides: Samba
# Required-Start: $network $local_fs $remote_fs
# Required-Stop: $network $local_fs $remote_fs
# Default-Start: 2 3 4 5
# Default-Stop: 0 1 6
# Should-Start: slapd cups
# Should-Stop: slapd cups
# Short-Description: start Samba SMB daemons (smbd/wsdd/nmbd)
### END INIT INFO
WSDD_DISABLED="Yes"
smbd=/usr/sbin/smbd
test -x "$smbd" || exit 0
nmbd=/usr/sbin/nmbd
test -x "$nmbd" || exit 0
wsdd=/usr/sbin/wsdd
test -x "$wsdd" && WSDD_DISABLED="No"
case "$1" in
start)
echo -n "Starting Samba: smbd"
$smbd
NMBD_DISABLED=`testparm -s --parameter-name='disable netbios' 2>/dev/null || true`
if [ "$NMBD_DISABLED" != "Yes" ]; then
echo -n " nmbd"
$nmbd
fi
if [ "$WSDD_DISABLED" != "Yes" ]; then
echo -n " wsdd"
start-stop-daemon --start --quiet --exec $wsdd
fi
echo "."
;;
stop)
echo -n "Stopping Samba: smbd"
start-stop-daemon --stop --quiet --pidfile /var/run/samba/smbd.pid
echo -n " nmbd"
start-stop-daemon --stop --quiet --pidfile /var/run/samba/nmbd.pid
echo -n " wsdd"
start-stop-daemon --stop --quiet --pidfile /var/run/wsdd.pid
echo "."
;;
reload|force-reload)
if [ "$WSDD_DISABLED" != "Yes" ]; then
start-stop-daemon --stop --quiet --pidfile /var/run/wsdd.pid
fi
start-stop-daemon --stop --quiet --signal 1 --exec $smbd
start-stop-daemon --stop --quiet --signal 1 --exec $nmbd
if [ "$WSDD_DISABLED" != "Yes" ]; then
sleep 1
start-stop-daemon --start --quiet --exec $wsdd
fi
;;
restart)
echo -n "Stopping Samba: smbd"
start-stop-daemon --stop --quiet --pidfile /var/run/samba/smbd.pid
echo -n " nmbd"
start-stop-daemon --stop --quiet --pidfile /var/run/samba/nmbd.pid
echo -n " wsdd"
start-stop-daemon --stop --quiet --pidfile /var/run/wsdd.pid
echo ""
echo -n "Waiting for samba processes to die off"
for i in 1 2 3 ;
do
sleep 1
echo -n "."
done
echo ""
echo -n "Starting Samba: smbd"
start-stop-daemon --start --quiet --exec $smbd
NMBD_DISABLED=`testparm -s --parameter-name='disable netbios' 2>/dev/null || true`
if [ "$NMBD_DISABLED" != "Yes" ]; then
echo -n " nmbd"
start-stop-daemon --start --quiet --exec $nmbd
fi
if [ "$WSDD_DISABLED" != "Yes" ]; then
echo -n " wsdd"
start-stop-daemon --start --quiet --exec $wsdd
fi
echo "."
;;
*)
echo "Usage: /etc/init.d/samba {start|stop|reload|restart|force-reload}"
exit 1
esac
exit 0
root@et10000:/etc/rc0.d#
Went back to 5.4.0.001 and no restores and using the wizard on lan 2 it states no internet connection,and vlan locked out until i go up and use remote to change it back to lan 2 and wondering if its the same on yours.
Back on with lan 2 no restores and guess what no et on browser scan
BrokenUnusableAccount
28-12-20, 21:51
@ccs
Oh I'd forgotten about wsdd.
I think it's just to help make samba visible to browse from Windows, possibly mainly Windows 10.
BrokenUnusableAccount
28-12-20, 22:46
@ccs
If you find nmbd isn't running it might be worth checking nmbd's log file, which on my Zgemma H7 is /var/log/samba/log.nmbd.
It's date/time and it's contents should help work out if nmbd has never started or if it started up and then crashed or halted for some reason.
root@zgemmah7:~#
root@zgemmah7:~#
root@zgemmah7:~# ps -AFH | grep -e "[sn]mbd\|wsd\d"
root 2061 1 0 7248 11092 0 21:41 ? 00:00:00 /usr/sbin/smbd
root 2065 2061 0 6600 3472 0 21:41 ? 00:00:00 /usr/sbin/smbd
root 2066 2061 0 6598 2260 0 21:41 ? 00:00:00 /usr/sbin/smbd
root 2068 1 0 4355 3340 0 21:41 ? 00:00:00 /usr/sbin/nmbd
root 2075 1 0 427 72 0 21:41 ? 00:00:00 /usr/sbin/wsdd
root@zgemmah7:~#
root@zgemmah7:~#
root@zgemmah7:~# ls -l /var/log/samba/log.nmbd
-rw-r--r-- 1 root root 544 Dec 28 21:42 /var/log/samba/log.nmbd
root@zgemmah7:~#
root@zgemmah7:~#
root@zgemmah7:~# cat /var/log/samba/log.nmbd
[2020/12/28 21:41:22.386397, 0] ../source3/nmbd/nmbd.c:907(main)
nmbd version 4.9.15 started.
Copyright Andrew Tridgell and the Samba Team 1992-2018
[2020/12/28 21:41:22.398134, 0] ../lib/util/become_daemon.c:135(daemon_ready)
daemon_ready: daemon 'nmbd' finished starting up and ready to serve connections
[2020/12/28 21:42:03.476868, 0] ../source3/nmbd/nmbd_become_lmb.c:397(become_local_master_stage2)
*****
Samba name server ZGEMMAH7 is now a local master browser for workgroup WORKGROUP on subnet 192.168.13.126
*****
root@zgemmah7:~#
root@zgemmah7:~#
@ccs
Oh I'd forgotten about wsdd.
I think it's just to help make samba visible to browse from Windows, possibly mainly Windows 10.
Thanks, wsdd is running on the working system, I can see the ET10K from windows 10, and I think I could earlier today when the network browser wasn't working.
.... having said that, it's also running on the ultimo4k, but windows can't see it. :eek:
BrokenUnusableAccount
29-12-20, 00:20
Back to square one, reflashed 5.4.003, no settings restore, start samba, no /usr/sbin/nmbd running.
Start /usr/sbin/nmbd by hand, network browser working fine.
So need to find out what autostarts /usr/sbin/nmbd (and now samba has been in the stopped state after 3 reboots, so probably part of the same rc file).
There's a startup script in /etc/rc0_to_6.d. I'm assuming they're all similar, named either K20samba or S20samba.
From those startup scripts it looks like nmbd wouldn't start if the samba configuration files had
disable netbios = yes
in them.
But as far as I can see on my Zgemma H7 only one of the sample files that aren't used has it, namely /etc/samba/distro/smb-secure.conf
Here we go (9 months self isolating does strange things to you) .... :)
/etc/samba after a clean install of 5.4.003, 68 packages installed with samba or smb in the name (samba not yet installed) ....
root@et10000:~# ls -lR /etc/samba
/etc/samba:
drwxr-xr-x 2 root root 528 Nov 11 12:24 distro
-rw-r--r-- 1 root root 20 Oct 8 05:12 lmhosts
drwxr-xr-x 2 root root 376 Dec 29 15:12 private
-rw-r--r-- 1 root root 744 Oct 8 05:12 smb-user.conf
-rw-r--r-- 1 root root 95 Oct 8 05:12 smb.conf
/etc/samba/distro:
-rw-r--r-- 1 root root 61 Oct 8 05:12 smb-branding.conf
-rw-r--r-- 1 root root 1457 Oct 8 05:12 smb-global.conf
-rw-r--r-- 1 root root 586 Oct 8 05:12 smb-shares.conf
lrwxrwxrwx 1 root root 13 Nov 11 12:24 smb-vmc.conf -> smb-vmc.samba
-rw-r--r-- 1 root root 29 Oct 8 05:12 smb-vmc.samba
/etc/samba/private:
-rw------- 1 root root 430080 Dec 29 15:12 secrets.tdb
-rw------- 1 root root 204 Dec 29 15:13 smbpasswd
-rw-r--r-- 1 root root 80 Oct 8 05:12 users.maps
After installing samba, 81 packages now installed (the same as a "working" samba), but /etc/samba has not changed and in particular, nmbd isn't running.
This is what "works" (quite different to the above) and is the same on an et10k and ultimo4k....
root@et10000:/tmp/etc# ls -lR samba
samba:
drwxr-xr-x 2 root root 180 Dec 29 14:38 distro
-rw-r--r-- 1 root root 20 Jan 27 2018 lmhosts
drwxr-xr-x 2 root root 120 Dec 29 14:38 private
-rw-r--r-- 1 root root 744 Jan 27 2018 smb-user.conf
-rw-r--r-- 1 root root 95 Jan 27 2018 smb.conf
samba/distro:
-rw-r--r-- 1 root root 168 Jan 27 2018 smb-branding.conf
-rw-r--r-- 1 root root 1189 Jan 27 2018 smb-global.conf
-rw-r--r-- 1 root root 253 Jan 27 2018 smb-insecure.conf (additional file)
-rw-r--r-- 1 root root 615 Jan 27 2018 smb-secure.conf (additional file)
-rw-r--r-- 1 root root 586 Jan 27 2018 smb-shares.conf
lrwxrwxrwx 1 root root 13 Dec 29 14:38 smb-vmc.conf -> smb-vmc.samba
-rw-r--r-- 1 root root 29 Jan 27 2018 smb-vmc.samba
samba/private:
-rw------- 1 root root 430080 Mar 20 2018 secrets.tdb
-rw------- 1 root root 204 Oct 13 09:40 smbpasswd
-rw-r--r-- 1 root root 28 Oct 26 2019 users-insecure.map
-rw-r--r-- 1 root root 80 Jan 27 2018 users.map
Something updated the above many years(2018?) ago but what is anybody's guess.
BrokenUnusableAccount
29-12-20, 20:25
@ccs
Hang on, the first listing is /etc/samba while the second seems to be /tmp/etc/samba which doesn't even seem to exist on my Zgemma H7.
@ccs
Hang on, the first listing is /etc/samba while the second seems to be /tmp/etc/samba which doesn't even seem to exist on my Zgemma H7.
I unpacked a settings backup of the et10k into /tmp to get hold of the "working" /etc/samba, just saved a bit of time. Just testing.:p
Another day, another idea. :) I wondered how I could possibly have got my ultimo4k working for the first time back in March.
The first image was 5.3.019, so I flashed it onto the et10k, no restores.
Gasp, /etc/samba is the "working" version (date stamps are a bit different from before because of my settings restores over the years) ....
root@vuultimo4k:~# ls -lR /etc/samba
/etc/samba:
drwxr-xr-x 2 root root 4096 Dec 30 10:40 distro
-rw-r--r-- 1 root root 20 Oct 26 2019 lmhosts
drwxr-xr-x 2 root root 4096 Dec 30 10:40 private
-rw-r--r-- 1 root root 744 Oct 26 2019 smb-user.conf
-rw-r--r-- 1 root root 95 Oct 26 2019 smb.conf
/etc/samba/distro:
-rw-r--r-- 1 root root 168 Oct 26 2019 smb-branding.conf
-rw-r--r-- 1 root root 1177 Oct 26 2019 smb-global.conf
-rw-r--r-- 1 root root 386 Oct 26 2019 smb-insecure.conf
-rw-r--r-- 1 root root 787 Oct 26 2019 smb-secure.conf
-rw-r--r-- 1 root root 586 Oct 26 2019 smb-shares.conf
lrwxrwxrwx 1 root root 13 Dec 30 10:40 smb-vmc.conf -> smb-vmc.samba
-rw-r--r-- 1 root root 29 Oct 26 2019 smb-vmc.samba
/etc/samba/private:
-rw------- 1 root root 204 Dec 30 10:43 smbpasswd
-rw-r--r-- 1 root root 28 Oct 26 2019 users-insecure.map
-rw-r--r-- 1 root root 80 Oct 26 2019 users.map
root@vuultimo4k:~#
What is really odd, everything I'd expect to be running after installing samba is already running, so all boxes see each other (and themselves).....
root@et10000:~# ps -AFH | grep -e "[sn]mbd\|wsd\d"
root 701 1 0 9519 8684 1 11:18 ? 00:00:00 /usr/sbin/smbd
root 705 701 0 8817 4144 1 11:18 ? 00:00:00 /usr/sbin/smbd
root 706 701 0 8815 3888 1 11:18 ? 00:00:00 /usr/sbin/smbd
root 709 1 0 6237 3884 1 11:18 ? 00:00:00 /usr/sbin/nmbd
root 718 1 0 498 1028 1 11:18 ? 00:00:00 /usr/sbin/wsdd
root@et10000:~#
Installing Utilities/Samba comes up with the green "running" box for an instant, followed by the invite to install packagegroup-base-smbfs-server.
That installs ok, but still needs an opkg update first.
So, Samba has changed since March, and affects both the et10k and ultimo4k (and probably everything else).
If I stop /etc/samba/distro/smb-insecure.conf from running by using the "non-working" version of smb-branding.conf thus...
root@et10000:/etc/samba/distro# cat smb-branding.conf
[global]
server string = OpenViX %h network services
root@et10000:/etc/samba/distro# cat smb-branding.conf.orig
[global]
server string = OpenViX %h network services
# include = /etc/samba/distro/smb-secure.conf
include = /etc/samba/distro/smb-insecure.conf
root@et10000:/etc/samba/distro#
I end up with these still running....
root@et10000:~# ps -AFH | grep -e "[sn]mbd\|wsd\d"
root 613 1 0 6235 3640 0 12:38 ? 00:00:00 /usr/sbin/nmbd
root 620 1 0 498 1016 0 12:38 ? 00:00:00 /usr/sbin/wsdd
root@et10000:~#
Not quite what I expected, as nmbd not running is part of the original problem.
Back to 5.4.003, before installing Samba,
root@et10000:~# ps -AFH | grep -e "[sn]mbd\|wsd\d"
finds nothing, which is what you'd expect, unlike 5.3.019 (post #64) which had everything running.
Commenting out the line "disable netbios = Yes" from /etc/samba/distro/smb-global.conf and installing Samba produces....
root@et10000:~# ps -AFH | grep -e "[sn]mbd\|wsd\d"
root 602 1 0 7933 7832 1 13:15 ? 00:00:00 /usr/sbin/smbd
root 606 602 0 7293 2872 0 13:15 ? 00:00:00 /usr/sbin/smbd
root 607 602 0 7291 2504 1 13:15 ? 00:00:00 /usr/sbin/smbd
root 609 1 0 4794 2888 1 13:15 ? 00:00:00 /usr/sbin/nmbd
root 617 1 0 499 1004 0 13:15 ? 00:00:00 /usr/sbin/wsdd
root@et10000:~#
which, I think, is what's wanted, with every box seeing what you'd expect.
I can find references to smb-global.conf on github, before netbios was disabled...
https://github.com/oe-alliance/oe-alliance-core/blob/e1a528415e6e243a2a6e9af17edc211b3b859b48/meta-oe/recipes-connectivity/samba/samba-3.6.25/smb-global.conf
and after
https://github.com/oe-alliance/oe-alliance-core/blob/4.4/meta-oe/recipes-connectivity/samba/samba/smb-global.conf
but no history, probably because of Samba version changes.
There are also references to the secure.conf and insecure.conf files here...
https://github.com/oe-alliance/oe-alliance-core/tree/e1a528415e6e243a2a6e9af17edc211b3b859b48/meta-oe/recipes-connectivity/samba/samba-3.6.25
Samba uses the directory /etc/samba to store its configuration files.
This directory comes as part of Vix, whether you use Samba or not.
My current version of Vix (5.4.003) is ok because I've had a working version for a number of years, and a settings restore overwrites
everything in /etc/samba. There is a fundamental problem here, because a new version of Samba was released a while ago, with
a new set of config files in /etc/samba. A system restore overwrites all these file, so I'm running the latest version of the samba packages
with an old set of config files.
Commenting out the line "disable netbios = Yes" from /etc/samba/distro/smb-global.conf works, but I guess is not the correct way to do it.
I'm assuming the file /etc/smb-user.conf should be edited instead, with the line "disable netbios = No", but I've not tried it yet.
BrokenUnusableAccount
31-12-20, 00:38
Samba uses the directory /etc/samba to store its configuration files.
This directory comes as part of Vix, whether you use Samba or not.
My current version of Vix (5.4.003) is ok because I've had a working version for a number of years, and a settings restore overwrites
everything in /etc/samba. There is a fundamental problem here, because a new version of Samba was released a while ago, with
a new set of config files in /etc/samba. A system restore overwrites all these file, so I'm running the latest version of the samba packages
with an old set of config files.
I think that probably doesn't matter.
Samba just looks at /etc/samba/smb.conf and also at any files mentioned in include lines in that first file.
Commenting out the line "disable netbios = Yes" from /etc/samba/distro/smb-global.conf works, but I guess is not the correct way to do it.
I'm assuming the file /etc/smb-user.conf should be edited instead, with the line "disable netbios = No", but I've not tried it yet.
Just bear in mind that smb.conf seems to be divided into sections, and the disable netbios line needs to be in a [Global] section.
I'm wondering if two boxs that both have "disable netbios = Yes" would be able to see each other.
Just bear in mind that smb.conf seems to be divided into sections, and the disable netbios line needs to be in a [Global] section.
The file that needs changing is /etc/samba/smb-user.conf, not smb.conf
This one line change works (at last) using the /etc/samba directory which comes with 5.4.003, no settings restore in sight, and nmbd starts ...
## User changes to Samba config go here
## You can overrule the insecure defaults and enforce password
## protected access to shares by uncommenting the following
## line, including the secure settings:
[global]
# include = /etc/samba/distro/smb-secure.conf
disable netbios = No
## You can hide and veto access to certain directories, e.g. mount points of
## other boxes in order to prevent round trips through your LAN
## In Samba 4.x this needs to be configured per share:
[Root]
# veto files = /ThisBox/OtherBox/AnotherBox/YetAnotherBox/
# hide files = /ThisBox/OtherBox/AnotherBox/YetAnotherBox/
[Harddisk]
# veto files = /ThisBox/OtherBox/AnotherBox/YetAnotherBox/
# hide files = /ThisBox/OtherBox/AnotherBox/YetAnotherBox/
I'm wondering if two boxs that both have "disable netbios = Yes" would be able to see each other.
With netbos disabled, nmdb doesn't run, boxes can only see boxes that have it running.
I think that probably doesn't matter.
Samba just looks at /etc/samba/smb.conf and also at any files mentioned in include lines in that first file.
The contents of the include files it was looking at before were different, but, in practice, added up to being very similar to the latest samba directory.
Unfortunately, my uniq/sort/diff attempts struggled to produce a definitive list of exceptions, but there were some.
If you get the dreaded "/media/autofs/..... file does not exist" when selecting a remote location in movielist,
have a look in /etc/auto.network If you find 2 entries for the failing location, delete them both while enigma2 is stopped.
Restart and then use network manager to delete the location, and then network browser to re-create.
One last observation. If you remove the samba service via network/utilities/samba then smbd, nmbd and wsdd no longer run, which is what you'd expect.
However, '/etc/samba' is not removed from the entry "config.backupmanager.backupdirs=" in /etc/enigma2/settings.
So /etc/samba will be in settings backups for evermore.
.
This means that if you have an old version of /etc/samba (like I had), removing the samba service, taking a settings backup, reflashing a new image followed by a settings restore won't have the desired effect, the old one will come back.
( '/etc/samba' is not in the settings entry "config.backupmanager.backupdirs=" after a clean flash of a new release, it is put there after you install samba for the very first time.)
.... and for anyone who thinks they may have come up with a workaround (removing the /etc/samba reference in settings for example), think again, because backupmanager puts it back in again! :):):)
So it looks like /etc/samba is backed up even if you've never installed samba.
But i have used an earlier image and setup from start and box is now seen,for the moment.
But i have used an earlier image and setup from start and box is now seen,for the moment.
Settings restores kind of fixed the problem by accident rather than design.
If you flashed an old image (you mentioned 2 years old a while ago) then it's very unlikely it needs "disable netbios = No" putting in smb-user.conf.
I reckon the new samba changes went in about 14 months ago.
The earlier images use an older version of samba which supports SMB1 and network browsing works fairly well and handles older windows shares (Win95/98/XP/NT/2000 etc). Win10 by default doesn't support SMB1, so the default now is SMB2 in most linux distros and network browsing doesn't work well. enigma builds now use SMB2 by default also. Most of my shares are on a linux mint (ubuntu derived) box which acts as a NAS. This too uses SMB2 by default. So I bit the bullet and set user and password on the various shares and access them by IP address/sharename and provide a userid and password in the mount manager in enigma and on my separate kodi box and it all works fine in the 5.4.nnn builds. It's all in the name of improved security, supposedly! There's no "one size fits all" option if you have a mixed MS Windows and linux network, particularly if you are running an older unsupported windows version (anything other than Win10). @SpaceRat is probably the best at describing the various scenarios and best settings, so if you search for posts by him you might get some good tips. I just gave up on browsing and access everything by IP address at this stage (but it can be done if you are prepared to keep configuring and maybe adopt the avahi method in an all linux environment). As @ccs has discovered, the samba config files are not being cleaned up properly when you remove samba from the system so that means you can end up with bits of old samba config and new samba executables. The older versions of samba had netbios enabled by default I think so that provided the network discovery, but that's a security issue also, so it's not the default now.
Powered by vBulletin® Version 4.2.5 Copyright © 2024 vBulletin Solutions Inc. All rights reserved.