Has anyone here ever ran a farm or cluster of servers/computers with slackware? How difficult did you find it installing & updating automatically? Or updating by hand if you prefer that too.

1 members found this post helpful.

icecream is already in current, so it's just a matter of setting it up. 1337_powerslacker also has a thread discussing icecream and Slackware. I also believe that Pat uses icecream for his build cluster for Slackware itself.

1 members found this post helpful.

I found a program called ClusterSSH that makes exact changes on multiple computers allowing binary updates quickly rather than by source for a cluster of computers. I don't know if there's something better than ClusterSSH.

1 members found this post helpful.

I use ClusterSSH to manage 20 workstations with Ubuntu (I usually run my own scripts) BUT I had problems a few times when I tried simultaneous configurations.

1 members found this post helpful.

Oops, I missed the part about managing those multiple systems.

This ServerFault question might be a handy exchange to look over and get many opinions.

1 members found this post helpful.

1.

Slackware GNU/Linux:
where lazy computer enthusiasts find refuge.

2.

Slackware GNU/Linux:
not reinventing the wheel since early 90s

3.

Slackware GNU/Linux:
we saw them all

3 members found this post helpful.

Yes, when I was the entire tech department for a startup called TSGS I ran a small web cluster and a small database cluster on Slackware, and at home I maintain a small dev cluster for personal use (was ten nodes once, now down to seven).

Those clusters were all small enough that I could use a custom install script from an external hard drive with a slackware gold copy on it, but nowadays I would use TAI to PXE-install larger batches of blank servers over the network.

For management, I have "mssh" which is a ssh multiplexer, but nowadays people use ansible to automate tasks. I also have some custom automation to keep things running smoothly -- whackamole is a security/countermeasures program similar to fail2ban, and unixadminbot detects and fixes a variety of minor problems (like directory permissions).

I have designed, implemented and operated large clusters professionally (2000 nodes at archive.org, 500 nodes at Discovery Mining, 20000 nodes at Autonomy/Zantaz) and would be comfortable running Slackware at similar scale.

6 members found this post helpful.

Had to stop using Slackware on my media player box because Sling refuses to work. I know it has nothing to do with systemd but every systemd based distro can stream fine while distros like Arch, Devuan, and all the other systemd free distros refuse to work.

** Absoutely spot on **

I first tried Slackware 8.1 back in the early Noughties. I just liked it as it got me doing things with my home PC's I'd never tried to do before.

I'd been on MSDOS and Windows since '88. My first foray into Linux lasted about 9 months then went back to Windows for work reasons until Oct 2016 when the NHS ransomware shennanigans made me realise Windows was a security joke and I was just a sitting duck.

First thoughts were Slackware again so got my 14.2 DVD set mailed to me for 50 bucks and set about learning Linux properly.

All I can say now is, "Why didn't I stick with it from 8.1?"

I never look back now, just forward.

1 members found this post helpful.

What I do won't scale very well; I'll just start off with that confession. When the pain gets too great, I'll look at doing things some other way. At my current work, they liked using Chef to upgrade infrastructure servers prior to jumping onto the Docker/Kubernetes bandwagon. I have read some Chef documentation in the past and was slightly frightened by what I perceived to be a hot mess was nonetheless considered a vast improvement over what was used before.

After grunting through the GNU package list, Swbis might be useful. (I haven't read much about it, so I really mean "might".)

EDIT: From Swbis documentation...

cough

At the moment, I have 5 Slackware systems. 3 are on Slackware64 14.2 and 2 (one of which is a VM) on Slackware64 -current.

As far as any of you are concerned, the 14.2 systems are named cranium, testbed, gateway. The -current systems are named currentbuild and sonicmaster.

My normal user has ssh keys on all the machines. testbed is a 14.2 slackbuilds server, while currentbuild is a -current slackbuilds server. I use either slackroll or slackpkg to keep my non-slackbuild packages up to date on all of my machines. testbed and currentbuild are as close to vanilla as they can be.

I use slapt-get to keep my slackbuilds packages up to date. Whenever I remotely log into a machine to upgrade it, I will launch a tmux session and run "su -" in that session prior to starting any upgrade commands.

2 members found this post helpful.