<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body>
<div class="moz-cite-prefix">On 23/02/2023 16:49, Martin Davies
G0HDB via Dxspider-support wrote:<br>
</div>
<blockquote type="cite"
cite="mid:63F798FE.31598.14D5BF25@marting0hdb.gmail.com">
<pre class="moz-quote-pre" wrap="">Hello all, I'm part-way through upgrading the GB7DXC-5 node, which currently runs
DXSpider 'Master' v1.55 build 249, to 'Mojo' v1.57 build 499.
I've installed the Lite version of 32-bit Rasbian 'Bullseye' (Debian 11) on a Raspberry Pi 3B
and have migrated the root filesystem to a USB-connected hard disk, and I've run the
SV5FRI script (thanks Yiannis!) to install the latest Mojo version of DXSpider onto the Pi.
The Pi is headless so everything is done using SSH access. The only problem that I seemed
to encounter during the DXSpider install was that for some reason the script didn't add the
sysop user to the 'sudoers' list so I had to do this manually later - I can now use sudo when
logged in as sysop.
As far as I can tell everything seems to be working as expected - the node, which I've
temporarily named GB7DXC-7 - appears to be running, I can log into it from a sysop console,
and when I (temporarily) connect the new node to my existing GB7DXC-5 node I can see the
DX spots coming through. So far so good!
However, there's a few aspects for which I need some help...
1. The node appears to start successfully whenever the Pi is rebooted - the SV5FRI
script seems to have created a file called dxspider.service in /usr/lib/systemd/system
but I'm unclear about whether anything else is needed to enable cluster.pl to be
restarted automatically after a crash or whatever. I'll be grateful for any clarification
on this aspect although I see that the dxspider.service file does include the statement
of 'Restart=always'. Is anything else needed?</pre>
</blockquote>
<br>
No.<br>
<br>
<blockquote type="cite"
cite="mid:63F798FE.31598.14D5BF25@marting0hdb.gmail.com">
<pre class="moz-quote-pre" wrap="">
2. I'll want to migrate as many user-related and other relevant files as possible from the
existing GB7DXC-5 node running 'Master' v1.55 build 249 over to the new node
running 'Mojo' v1.57 build 499. Which are the essential files that I'll need to copy
across from the old node to the new node?</pre>
</blockquote>
<br>
Hmm... I probably need to have a look at that script. My standard
recommendation for UPGRADING to mojo from master is to start from an
existing installation and that means having started with your step 3
first. But, using the power of git, all is not lost.<br>
<br>
Firstly do a 'systemctl stop dxspider' or 'service dxspider stop'
the first will work on buster and the second should work (otherwise
you may need to install a package 'apt install
init-system-helpers'). <br>
<br>
cd /spider/local; cp DXVars.pm /tmp/DXVars.pm.dxc7<br>
rm -rf /spider/local_data/*<br>
<br>
Do your step 3. After which you will have a copy of the complete
tree from your old machine on your new one. <br>
<br>
cd /spider/local_cmd; mv crontab crontab.save<br>
<br>
cd /spider<br>
git reset --hard<br>
git branch<br>
<br>
check to see that there is an '*' in front of the mojo branch name.<i>
</i>If not then do a 'git checkout mojo'.<br>
<br>
At this point, you have a choice: you can just expect it to 'work'
(minus crontab stuff) or you can test it again with DXC-7 if you
want to do that (I wouldn't bother) then do these steps:<br>
<br>
cd local; mv DXVars.pm DXVars.pm.dxc<br>
cp /tmp/DXVars.pm.save DXVars.pm<br>
/spider/perl/update_sysop.pl<br>
<br>
This command, as a side effect, will do the users file conversion. <br>
<br>
You will now have a working DXC-7 node to test. Run it manually
first as below. When you are happy that it is working as desired you
will need to shutdown the node (either 'shut' in a console.pl or
just <ctrl-c> it in the shell). To now run as DXC(not-7) then:<br>
<br>
cd /spider/local; cp DXVars.pm.dxc7 DXVars.pm <br>
<br>
And, assuming you have removed all the extraneous stuff as
recommended in UPGRADE.mojo, just change $mycall to 'GB7DXC' and
$myalias to your taste.<br>
<br>
/spider/perl/update_sysop.pl<br>
<br>
Run it manually first:<br>
<br>
cd /spider/perl/; ./cluster.pl<br>
<br>
Check that it all works. This is the point at which you stop BOTH
nodes, do any IP address changes you require, and (on the new node)
restart the whole machine. The cluster.pl should be running. <br>
<br>
If it is, then:<br>
<br>
Start watchdbg in another shell and:<br>
<br>
cd /spider/local_cmd; mv crontab.save crontab<br>
<br>
You should see all your node connects all spring into life over the
next few minutes.<br>
<br>
Do not switch your old machine back on :-)<br>
<br>
<blockquote type="cite"
cite="mid:63F798FE.31598.14D5BF25@marting0hdb.gmail.com">
<pre class="moz-quote-pre" wrap="">3. I assume I'll need to use rsync to copy the relevant files across from the old node to
the new node; I'm largely unfamiliar with using rsync for this purpose so any guidance
on the syntax to use will be much appreciated! I find the documentation for rsync to
be rather confusing for my limited Linux skills and experience!</pre>
</blockquote>
<br>
on your new machine: 'rsync -avz
sysop@<oldmachinename>:/spider/* /spider --exclude .git' <br>
<br>
NOTE: you may need to install the rsync package on one or both
machines if you have not used them before.<br>
<br>
<blockquote type="cite"
cite="mid:63F798FE.31598.14D5BF25@marting0hdb.gmail.com">
<pre class="moz-quote-pre" wrap="">
Many thanks in advance for any help and suggestions that might be offered.
73
---
Martin, G0HDB
_______________________________________________
Dxspider-support mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Dxspider-support@tobit.co.uk">Dxspider-support@tobit.co.uk</a>
<a class="moz-txt-link-freetext" href="https://mailman.tobit.co.uk/mailman/listinfo/dxspider-support">https://mailman.tobit.co.uk/mailman/listinfo/dxspider-support</a>
</pre>
</blockquote>
<br>
</body>
</html>