Hey guys,

I had some bad luck today when I was trying to fix an account that had trouble with FTP. I decided to remove the user and add him again and reset all configuration. Anyhow, to make things short, I accidently typed rm /* -R -f and without looking hit the enter button, as soon as I realised what I wrote I hit ctrl+c. Too bad the /bin/ folder was gone by this time and standard commands like ls didn't execute anymore.

My question is if there is any way to recover these files by a system repair or something? The server is used to host a heavy loaded site which can't afford any downtime, and silly me didn't make a backup of the whole HDD, just the important folders (not the system files).

Currently I don't have the balls to restart the server as I know this will probably turn into a dissaster. I also don't have straight access to this server since it is located in a datacenter (I can go there if absolutely necessary but I rather don't).

Anyone has some advice? I'd be glad to hear if there's a command to repair/redownload system files!

Kind regards,
Dennis Bronk.

P.S. sorry if the post is a little hard to read, I am dutch.

What distribution+release? If it has proper and mature package management can you list or verify package contents to see which are affected?

I am running CentOS 5.

Currently the situation has gotten worse, I was having dinner and when I returned my SSH connection was closed and WinSCP too. I can't connect to any of these anymore, the server just abrupetly closes the connection.

Websites are still up though.

Through some workarounds I can still have limited control over the server. I am running DirectAdmin which enables me to edit files, and I could also execute CGI scripts.

This is bad ...

edit:
I think this requires a reinstall of linux, I am waiting for a few hours until most customers won't order anything and I'll make a backup of MySQL database and the website, put it offline immediatly afterwards and temporarily host it on a Dell Poweredge R200 that I use for testing purposes.

If anyone has some advice though, or knows a solution I'd be glad to hear since I really don't want to go through this kind of trouble. I am probably able to run executables through scripts and have full control over configuration files, so I still have a little control.

Usually you would be able to list packages from the RPMDB that have items in /bin, but since 'rpm' resides in /bin as well you would need to copy /var/lib/rpm over to another machine to read it with something like 'rpm -qa --dump|awk '/^\/bin/ {print $1}'|xargs -iX rpm -qf 'X'|sort -u'. The resulting core packages should be a list like "ash bash coreutils cpio dbus ed filesystem (fuse) gawk gettext grep gzip initscripts iputils kbd keyutils mailx mktemp net-tools procps redhat-lsb rpm sed setserial SysVinit tar util-linux vim-minimal". Since /bin/rpm is gone you'll have to curl/wget these packages from the std repo to your staging/rescue Centos machine, unpack each with 'cd /tmp; rpm2cpio "${packagename}" | cpio -idmv', then copy the ./bin parts over to the victim system.

Ok, that doesn't sound like a hard job. Somewhere on the internet I saw an advice on copy'ing /usr/bin to /bin/ as a backup replacement. But I can't imagine those folders are the same ....

I am lucky only my bin got deleted (atleast, that's what I hope, no other folders dissappeared but some files may have been deleted).

Such a small typo, such a great mess it caused heh. I currently don't have a backup linux distro running, but I'll get one up ASAP using dual boot and a spare harddisk I have for my own PC.

Meanwhile I'll get the important files and I'll try to find out if I can write to / using scripts. I think that's going to be a though one too. SFTP is not working anymore, though normal FTP is.