Initial installation works fine and in last modules got this
also can you please configure script to skip not so important steps just like this one if something not working to the next steps as virtualmin installation stops execution where-ever it gets any error.
[21/23] Configuring Fail2banFirewalld ▒▒▒▒▒▒▒Error: Systemd service fail2ban cannot be created unless a command is given
▒▒▒▒▒▒▒Error
Systemd service fail2ban cannot be created unless a command is given
▣▣▣ Cleaning up
[WARNING] The following errors occurred during installation:
◉ Postinstall configuration returned an error.
Tried fixes given in older issues but all of them are already there in cgi files.
There are many other CLI arguments which can be used with the script.
--help|-h Print this summary of options and exit
--list-bundles|-s List available installation bundles
--list-plugins|-p List available plugins
--bundle|-b A bundle of plugins to execute
--log|-l Path to a file for logging actions
--include|-i One or more extra plugins to include
--exclude|-x One or more plugins to exclude
--test|-t Test services after configured (when available)
Never mind thank you all for you suggestions, the issue was maybe with the VPS server as the kernel version 4.18.0-25-generic as when tried on aws server with kernel verison 5.3.0-1032-aws everything went fine.
Although I manually configured postfix config, but to be double sure ran configuration again for last 4 steps on vps server with kernel version 4.18.0-25-generic
Great. I’ll fix the bug causing the “uninitialized value” errors…it’s happening because it’s not being run as part of a bundle (which should work fine, and the error is harmless in this case).
Great. I’ll fix the bug causing the “uninitialized value” errors…it’s happening because it’s not being run as part of a bundle (which should work fine, and the error is harmless in this case).
@Joe, besides, could we simply display a red tick, with small error messages underneath, on failure, instead of just stopping the process completely?
No. If a step is missed, it may cause subtle problems in other steps later. An install that fails at any step (other than warnings) should not be used until the install is completed. Users will ignore any errors. I don’t know why, but they will.
I need to make it more insistent when something fails. It needs to be loud and it needs to stop and clearly say it stopped and didn’t finish. As it is, several times a week someone comes in with weird problems caused by an incomplete installation, and they’re doing a bunch of stuff manually to fix what the installer didn’t finish. So…the solution is more insistent and loud failures, not rolling on past them as if nothing happened…people will absolutely ignore any error message that is not the very last thing they see (and, some people will still ignore it, somehow).
I don’t have time to dig into it right now, but the error detection, I think, needs to also include notifying the user the number of steps that didn’t complete. Again, we need to very plainly spell it out, because e.g. when users see 20/23 as the last step, somehow they think the install is fine, even though it definitely is not. So, I guess the script needs to say something like “3 of 23 steps did not complete. Your installation is not complete. Please check the troubleshooting guide at X or post to forum.virtualmin.com blah blah blah for help.”
And, the solution to a step that commonly fails is to make the step more robust. I haven’t looked into why this one failed (again, no free time), but we just need to accommodate all of the common failure modes…and for uncommon failure modes we need a clear indication that something went wrong and the user needs to do something to resolve it.
I don’t know that this is a common failure, though, so maybe it’s a moot point, in this case.
