After encoutering numerous problems in setting up SANED network access with systemd on a Linux Mint 18.3 , here are some important points

1) make sure you do not have saned configured to run under inetd or xinetd as might be the case from an upgraded installation

2) systemd needs a file for socket and an instance service file. To avoid loss of customizations with package upgrades, put these in /etc/systemd/system not overwrite those in /lib/systemd/system

#*****************************************************************************#
#|
#|  file : /etc/systemd/system/saned.socket
#|
#*---------------------------------------------------------------------------*#

[Unit]
Description=SANED network daemon activation socket

[Socket]
Accept=yes
ListenStream=6566
MaxConnections=1

[Install]
WantedBy=sockets.target

#*****************************************************************************#

The second item it needs is an instance service file NOT a service file. This means that the file name contains an "@" as in saned@.service and the contents

#*****************************************************************************#
#|
#|  file : /etc/systemd/system/saned@.service
#|
#*---------------------------------------------------------------------------*#

[Unit]
Description=SANE network daemon instance %i
Documentation=man:saned(8)
After=local-fs.target network-online.target
Requires=saned.socket

[Service]
Environment=SANE_CONFIG_DIR=/etc/sane.d
ExecStart=/usr/sbin/saned
Group=saned
User=saned
StandardInput=null
StandardOutput=syslog
StandardError=syslog

[Install]
Also=saned.socket

#*****************************************************************************#

If you want to do debugging you can add additional Environment lines

Environment=SANE_DEBUG_DLL=255
Environment=SANE_DEBUG_NET=255

For completeness, to mirror the setup in /lib/systemd/system and to emphasize the point when looking in /etc/systemd/system that the file is not missing and that saned@.service is not misnamed, symbolically link /dev/null to /etc/lib/systemd/system/saned.service

 /etc/systemd/system/saned.service -> /dev/null

This is unecessary because of the link present in /lib/systemd/system, but it makes it clear when looking at the /etc/systemd/system directory the configuration being used.

Do a systemctl enable of saned.socket and it will create a symbolic link under the socket.target.wants directory

/etc/services should have the entry

 sane-port       6566/tcp   sane saned

Ensure that /proc/sys/net/ipv6/bindv6only is set to 0 (and not set to 1 due to some hack from an old bindipv6only.conf file /etc/sysctl.conf.d) if you are wanting network connections on IPv4 to work.

It may be necessary to ensure that the "net" featured is turned on in /etc/sane.d/dll.d configuration file for your scanner if it is some non-standard configuration.

In your /etc/sane.d/saned.conf ensure your have a "localhost" entry -- 127.0.0.1 should also work but when saned starts up it says checking for localhost, so this is to be consistent. Then add the host names or network IP address ranges permitted to access the service.

Adjust your firewall rules if necessary.

You do not need to add the host IP address to the /etc/sane.d/net.conf file, this will result in you being offered both a local and a network connection to the scanner from the host, so keep it simple for the server host, but do add the server host IP address or name to the net.conf file on the client hosts which need to access the service.

If you then do a systemctl start saned.socket followed by a systemctl status saned.socket you should see

 saned.socket - SANED network daemon activation socket
   Loaded: loaded (/etc/systemd/system/saned.socket; enabled; vendor preset: enabled)
   Active: active (listening) since Wed 2019-05-29 20:12:40 BST; 30min ago
   Listen: [::]:6566 (Stream)
 Accepted: 21; Connected: 0

Now where the "magic" comes in, is that when a connection is made on the socket, systemd fires up an instance of the saned service using the socket name as the instance name (which is why the standard input is set to "null" and NOT "socket" in the saned@.service file).

So if when a remote connection is made you do

systemctl -all -l --no-pager | grep saned you should see

  saned@21-192.168.21.12:6566-192.168.11.12:49314.service                                    loaded    active   running   SANE network daemon instance 3 (192.168.11.12:49314)
  system-saned.slice                                                                        loaded    active   active    system-saned.slice
  saned.socket                                                                              loaded    active   listening SANED network daemon activation socket

where the instance number increases by one for each connection.

If you want to advertise the service via Avahi, you could add a service file under /etc/avahi/servvices

<?xml version="1.0" standalone='no'?>
<!DOCTYPE service-group SYSTEM "avahi-service.dtd">

<!-- #********************************************************************# -->
<!-- #|                                                                  |# -->
<!-- #|  file : /etc/avahi/services/saned.service                        |# -->
<!-- #|                                                                  |# -->
<!-- #|__________________________________________________________________|# -->

<service-group>

     <name replace-wildcards="yes">%h.example.COM Network Scanning</name>

        <service>
            <domain-name>local</domain-name>
            <host-name>server_name.local</host-name>
            <port>6566</port>
            <type>_scanner._tcp</type>
        </service>

  </service-group>

 <!-- #********************************************************************# -->

replacing example.COM and server_name as appropriate. The value for "type" was taken from

 http://www.dns-sd.org/ServiceTypes.html

If you set up [MD-raid> LUKS > LVM > {pool/swap, pool/root etc}] stack then you get one PW prompt and suspend to disk still works...

br: a random guy

After everything is submitted, I made some changes to my commits.

When I run arc diff for a particular commit, the process is really confusing. For example, with arc diff --update DXXX, I get prompted with the whole list of revisions...

The -c option is a great recommendation, but I've been trying out https://github.com/StanfordSNR/guardian-agent and I like it even better; it gives you much more information about what is happening: which computer is asking for permission, which key they want to use, what server they're going to connect to, and what command they want to run using it. You can make a much more informed decision, and you can save those decisions so that you only have to decide for novel situations.

Also, the ProxyJump command is much nicer than ProxyCommand, but also newer. It's easier to use and harder to misuse.

Thanks, works in ubuntu server 18.04.1 ! but i need to set the md device --readwrite, too.

(this "feature" is unfixed for over 9 years )

I have follow the guid to do in debian 9 , everything works perfectly , thanks ! but in the modify /tmp/mntroot/etc/fstb section, debian 9 use UUID instead of sda , and everything rest were almost the same as the guide .

I've follow this tutorial, everything went OK but grub. This part was difficult because of two reasons: 1) I use the GPT and the error was that boot partition doesn't have the bios_grub flag - don't forget to assign when create partitions on the new disk
2) the new 16.04 system has different raid modules to be loaded at grub start like dm_raid, megaraid etc. This part should be updated.

I didn't need to install lvm2, as it was on my unbootable system. I also had some minor partition/volume differences.

My issue is documented at the Ubuntu forums

That all said, I did have a major issue with DNS resolution not functioning after this was done. I'm wondering if "update-initramfs" lead to this issue specifically (I made other changes I can't recall clearly).

If other experience loss of DNS via systemd.resolved failure, please note it here and on my post in the Ubuntu Forums. My fix is listed there, although I'm effectively disabling systemd.resolved.

Hello,

After some searching online I tried to update the BIOS on an Ideapad 120s via USB bootable stick with DOS.

How to flash via DOS boot usb.........

1. First you need Rufus and an 1GB USB stick - what I had for a stick

01a. create a usb dos boot disk with rufus

1. You need the BIOS, ex 6gcn25ww.exe

2. on an Linux machine you use innoextract to extract the executable fron the auto-unpacking archive. will have same name and extension

3. you put the newly obtained exec in the root of the dos boot usb stick

4. boot the machine with said usb stick

5. type the name of .exe file at the promt, wait for extract process, wait for reboot, let the exe do its job - a new boot window will appear with the progress bar.....wait some more.....wait til machine reboots......check new bios......success !

finish

My 2 cents for people with noob skills on Linux.

I hope it will help someone....

Best regards