linux novice common mistakes.

[Khyron320]

I have a job working with linux everyday. What are some silly mistakes you have made as a novice user to linux?

I keep calling scp "supercopy" in front of the other techs. Its actually "securecopy". Old habbits die hard I guess.

Another one is proper use of grep. I somehow formed the habbit of doing "cat file.log | grep error" instead of proper use "grep file.log error"

So what are some of yours?

 

[postaled]

I wouldn't really call this a common mistake but one of the first things I wanted to know in gnome was how to add a launcher that would open up a terminal and run a command like ssh or ftp


Figured this may save some people some time.

just create a launcher and put the command as

gnome-terminal -e 'ssh postaled@postaled.com'


that would make a launcher in gnome to open up a terminal to ssh the user postaled to postaled.com



Thats my two cents.

 

[kedlav]

vi hugefile

 

[Aknosis]

rm -rf /

 

[Khyron320]

Oh come on we are all waiting for the dd story =D

 

[LuckyTaxi]

Let's see, what have I done in the past 5 yrs.

- Make sure LUN isn't being used by a server. Document your sh!t. I had a colleague wipe out a LUN that attached to one of our oracle servers.

- For recursive, it's "scp -r" and "cp -R" ... take note of the capital and lowercase "r"

- rsync used to confuse the hell out of me. A trailing slash on a directory means to copy the content inside the directory. No slash means copy everything inside the directory along with the parent folder. (i.e. rsync -avz myfolder/ <target> vs rsync -avz myfolder <target>). The latter will create a myfolder on the target while the former "copies" everything inside of "myfolder." Things can get messy if you don't pay attention.

- I modified my bash prompt to show current working directory. You dont know how many times I did an rm -rf * in the wrong directory.

 

[degibson]

Are we talking linux users, or linux admins, or both?
Users: "How do I start Excel?"
Admin: "How do I start Excel?"

One is way worse than the other.

That said, here's another bad mistake:

Enter new password for root: root
Enter password again: root

 

[Nothinman]

Another one is proper use of grep. I somehow formed the habbit of doing "cat file.log | grep error" instead of proper use "grep file.log error

I wouldn't call that an error, doing it that way lets you change the pattern easily by just hitting Ctrl+W. Using just grep may be more work since the pattern goes before the filename.

My biggest mistakes are typing a command on the wrong server, with so many screen windows open it's very easy to forget which box I'm on.

 

[Gamingphreek]

Well I wasn't a Linux noob when I did this, but it was early and I wasn't thinking....

'sudo find / '[filename]' | rm -rf *.*'

I was trying to search from the root directory and find a certain group of filenames/folders which I would then delete. Obviously, outside of the fact that my find command was horribly incorrect, the entire concept was a terrible idea to combine into 1 command (ESPECIALLY with sudo).

I then watched in horror as my system systematically deleted its entire contents (including my boot loader). As icons and menu bars disappeared, the recursive delete finally stopped when it deleted its own binary file.

Not common, not a noob move, and certainly one of the dumbest things I have ever done on my computer

-Kevin

 

[sourceninja]

My favorite mistakes is when iChat takes over my typing when I'm in the middle of something and not really paying attention.

User: Hey, what's up?
Me: sudo /etc/init.d/bind9 restart
User: Ummm what?

 

[Khyron320]

Here is a biggy I caught yesterday. I was looking for a command in history. Somebody had accidentaly typed their password on a shared user account and didn't even bother to wipe it out of .bash_history

 

[degibson]

vi hugefile

lpr hugefile

 

[zoiks]

# chmod -R 755 /

Needless to say that screwed everything up. I was trying to type in /opt/... but by mistake I hit enter right after the first slash.

We had a sys admin who deleted our vm that hosted our bug tracking system. We (the whole company) salvaged everything that we could from our browser caches etc and reconstructed the tracking pages. The sys admin was let go soon after.

 

[JD50]

# chmod -R 755 /

Needless to say that screwed everything up. I was trying to type in /opt/... but by mistake I hit enter right after the first slash.

We had a sys admin who deleted our vm that hosted our bug tracking system. We (the whole company) salvaged everything that we could from our browser caches etc and reconstructed the tracking pages. The sys admin was let go soon after.

No backups?

 

[degibson]

No backups?

A screwup in itself.

 

[electroju]

Setting tmpfs size to 100% of RAM or greater and bind /tmp to tmpfs. Then compile something big or forget to empty /tmp. Soon enough the computer will crawl swapping pages from memory and into swap.

Do not make /var/log/messages be bigger than the actual throughput of the hard drive or else your setup will be slow.

Be careful with setuid.

Do not use GUI programs to help configure daemons or programs that have config files formatted in plain text. Better, do not use GUI at all for administrating tasks. GUIs are not administrators best friends.

It is directories, not folders for crying out loud.

 

[zoiks]

No backups?

Nopes. That's one of the reasons he was let go.

 

[Fallen Kell]

Nopes. That's one of the reasons he was let go.

I guess the more important question is, was he given a budget to make the backups? We find systems all the time that are not backed up, but that is because when the system admin requested the money needed to add the system into the existing network backup solution there was never money for that, or if it was for tapes, there was never money for that either... I see lots of penny wise and dollar stupid decisions all the time. It is almost always about the short term savings, with no one ever looking at the long term implications.

 

[Nothinman]

I guess the more important question is, was he given a budget to make the backups? We find systems all the time that are not backed up, but that is because when the system admin requested the money needed to add the system into the existing network backup solution there was never money for that, or if it was for tapes, there was never money for that either... I see lots of penny wise and dollar stupid decisions all the time. It is almost always about the short term savings, with no one ever looking at the long term implications.

True, but on the other side of things I don't think I've seen a SMB with a proper backup/restore plan even if they've paid for crap like BE. Usually it's setup poorly and forgotten about until they need something and then they realize something was missed, backups have been failing, etc.

 

[RalphTheCow]

I just recently tried to install Linux after the Windows partition on a 160 GB drive on a system that originally came with a 40 GB disc, so the BIOS was unable to boot from a partition that far out - I then found there are 8 GB and 137 GB limits for old BIOSes.

Both Ubuntu and Mint installed just fine, and then both booted to a very unfriendly "no such partition" message and a grub recover prompt. So I was instantly sent back to a time to when Linuxes were not so user-friendly.