/tmp directory is mounted noexec. Installation cannot continue.

Howdy -- yeah, your fstab file looks okay.

What output do you receive if you run the command "mount"?

Aha! So according to this line, /tmp is indeed mounted as noexec:

/tmp on /tmp type none (rw,noexec,nosuid,bind)

Now, your fstab is correct... so something else is causing that to occur.

Are you by chance using a VPS of some sort? If so, it's possible that's being automatically done by your VPS provider.

In the meantime though -- running this command will remount your /tmp without the noexec flag:

mount -o remount,exec /tmp

Can you paste in the output of "mount" after running the command "mount -o remount,exec /tmp"?

Hmm, it actually looks like you have /tmp mounted twice. I overlooked this in your earlier output, but you had these two mount entries:

However, neither of those are in your fstab file, so once you reboot -- that should all be reset, in theory.

If your server isn't live, I might actually suggest rebooting, and seeing what your mount output looks like at that point... as it appears as if those issues will all correct themselves at that point.

Since /tmp isn't being set as noexec in your fstab file, Virtualmin should be able to install at that point.

So that's the output of "mount" after rebooting your system?

Can you verify that your fstab hasn't changed at all, by pasting in the output of the command "cat /etc/fstab"?

However, if it hasn't changed -- that means something on your system is mounting /tmp outside of your fstab. That doesn't happen on a default installation of CentOS -- do you have any idea where that might be setup?

Okay, you're setup is atypical, but we can make that work :-)

My suggestion is to run this command not once, but twice:

umount /tmp

After running that command twice, try running the Virtualmin installer again. It should complete successfully at that point.

Once the Virtualmin installer completes, go ahead and reboot, and that'll reset all your /tmp settings again.

Make sure your shell's current directory is not /tmp when running umount .

Also, you can check for programs using /tmp with the command fuser -m /tmp

Heh, that's quite a pickle :-)

In order to unmount /tmp, you'd need to stop any running process that has a file open in /tmp.

However, that's quite a few of them. So, your options are --

Kill all of those running processes, and when "fuser -m /tmp" no longer shows any processes holding a file in /tmp open, you can then umount /tmp (which you'd run twice).

Alternatively -- the setup your provider gave you is fairly atypical... you may want to ask them to help you disable it, at least temporarily. That is, you could request that they turn off the "noexec" option for /tmp. Since that's not being enabled from within /etc/fstab, only they would know how that's configured.

Once they turn off the noexec option on /tmp, the installer should work just fine.

Well, just remember that the support area here is typically for Virtualmin Pro customers. GPL users would generally want to use the Forums for obtaining support :-)

However, you can use the "kill" command from the command line to kill off those processes.

Hmm, so is WHM/cPanel installed on your system? That may be a problem for two reasons.

One, because having two different control panels on the same system could cause a conflict with each other, as each has their own way of doing things.

Second -- in some Googling, I see that the way your /tmp dir is setup may be part of a cPanel configuration.

If you do indeed have WHM/cPanel installed, is it possible to ask your provider for a fresh OS installation, with no control panel?

The Virtualmin installer works best on a fresh install of your distribution, prior to any other control panels being installed.

Howdy -- so is your server live at the moment, serving websites, email, or such?

Also, can you run the command "ps auxw", save the output do a .txt file, and attach that to this request?

Lastly -- Virtualmin is a tool for managing shared web hosting accounts.

If you're interested in managing VPS's -- Cloudmin may be what you're after. Cloudmin GPL can manage the free Xen and LXC based VPS's. Cloudmin Professional can manage a number of additional VPS types.

You can find more information on Cloudmin here:

http://www.virtualmin.com/documentation/cloudmin

The loop seems to be here - this was a parent that had a child or visa versa - its all coming back to me - when we move the child to the parent level, this problem started.

How can I fix, expecially if I cant access virtualmin yet

Can I rename/move/delete /etc/webmin/virtual-server/domains/141444784422529 to elimninate its confusion?

virt6already= edit_delete=0 hashpass=1 edit_passwd=1 db=xxxxxxxxlaundromat ugid=594 virtualmin-slavedns=0 proxy_pass_mode=0 edit_redirect=0 plan=12345678901234567890 webmin=0 dom=xxxxxxxxlaundromat.com edit_admins=0 uid=1077 name=1 virtualmin-awstats=0 postgres=0 web_port=80 virtualmin-dav=0 gid=594 virtualmin-git=0 source=domain_setup.cgi virtualmin-signup=0 subprefix= edit_sharedips=0 ip=108.163.170.226 md5_enc_pass=xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx dir=0 web=1 group=xxxxxxxx lastsave=1520027726 netmask= edit_phpver=0 edit_dnsip=0 home=/home/xxxxxxxx/domains/xxxxxxxxlaundromat.com/domains/xxxxxxxxlaundromat.com virt=0 edit_aliases=1 netmask6= crypt_enc_pass=xxxxxxxxxxxxxxxx edit_spam=0 limit_dir=1 ugroup=xxxxxxxx ip6= edit_scripts=0 digest_enc_pass=xxxxxxxxxxxxxxxxxxxxxxxxxxxx proxy_pass= edit_restore=0 subdom= virtualmin-registrar=0 edit_dbs=0 dns=1 creator=wcr db_mysql= edit_phpmode=0 virt6=0 mysql_enc_pass= emailto=xxxxxxxx@xxxxxxxxlaundry.com edit_spf=0 user=xxxxxxxx

CapstoneWorks , we're very happy to help with the issue you're seeing, but it would be best to troubleshoot it on the existing support request you have open, rather than on some of these older requests.

I'm going to work with Joe and Jamie on your issue, and we'll need to keep all the info in one place so we can review what's going on and help get that sorted out. Thanks!