Linux question of the day

W Stewart

Guess I was wrong after looking it up. I'll leave it for somebody else to take a shot at it.

UnixGuy

I wrote a short blog post about atime, ctime, and mtime in 2012:
Linux and UNIX: Understanding and modifiying File timestamps | Sahara Geeks

W Stewart

Here's one that you'll probably come across early on in your career if you work with linux.

I changed out a motherboard that has an integrated NIC. The network interface now says eth1 every time the system boots instead of eth0.

1. Why does it do that?
2. How can I fix it?
3. Bonus if you can give me at least two alternative ways to fix the problem other than the right way.

You can even do some googling for number 3 if you need to.

Expect

70-net-persistent file under udev is responsible for NIC mapping

W Stewart

^ that's pretty much what I was looking for. You could remove the file or just change the mac address in the file to the mac address of the new NIC. Any alternative ways to name the interface?

W Stewart

Just bough a decommissioned server at work and while setting everything up I remembered a question that somebody asked about determining if a device is hardware or software raid. Just wanted to point out that at least with an lsi scsi controller you can run

smartctl -d auto /dev/sda

to determine if it's a raid drive so long as the device supports smart monitoring.

Also to the question I posted above, you can rename an interface by creating the /etc/iftab file or by loading the NIC driver with an alias in the modprobe.conf file such as

alias <nic driver> <eth3>

W Stewart

free -m

             total       used       free     shared    buffers     cached
Mem:          3832       3388        444        135        177       1557
-/+ buffers/cache:       1653       2179
Swap:         3971          0       3971

Question of the day. My server has 444 megs of ram available for any applications that need it. True of false and why?

finhack

false.

it has 2179mb free left for usage. the 444 does not include memory in buffer/cache which linux allocates to programs and releases when new applications rise.

W Stewart

Correct, bonus questions that may require some experience with a linux system. What happens when I run out of memory(including memory used for disk caching and buffers) and swap space.

paul78

hmm - I assume the kernel starts to reclaim non-dirty pages but if you are toast then the oom killer starts to kick in.

finhack

I have an interesting one!

how to recover from chmod -x /bin/chmod ? (Real question, no joke)

W Stewart

paul78 wrote: »

hmm - I assume the kernel starts to reclaim non-dirty pages but if you are toast then the oom killer starts to kick in.

correct

W Stewart

finhack wrote: »

I have an interesting one!

how to recover from chmod -x /bin/chmod ? (Real question, no joke)

As root, run

setfacl -m u::rwx; setfacl -m g::rx; setfacl -m o::rx

You could also boot to a live cd and use the chmod command from the live cd to change permissions.

paul78

finhack wrote: »

how to recover from chmod -x /bin/chmod ?

Great question! I really liked this one.

@W Stewart - good answer - I wasn't familiar with that command.

What's neat about this question is that I bet there are several ways to do it depending on your background. The 2 solutions that I thought of is not as simple as using setfacl.

I am wondering if there are shell built-ins that can be used to solve this as well.

My own idea was the following:

1) Write your own quick app to turn on the execute bit

#include <sys/stat.h>

main()
{
  chmod ("/bin/chmod", S_IRWXU|S_IRGRP|S_IXGRP);  
}

That will turn on the execute bit and then you can use /bin/chmod.

or

2) Call the ELF loader directly on /bin/chmod to fix the execute bit. Something like this:

/lib/ld-linux-x86-64.so.2 /bin/chmod +x /bin/chmod

finhack

nice answers,

another one:

What would this command do? (and no, don't try it on production)


:(){ :|: & };:

W Stewart

That's a fork bomb. It will run a function called : which calls itself and then pipes the output into another call of itself and runs in the background. It will keep spawning more recursive functions until every instance is killed, the system is rebooted or the system runs out of resources.

W Stewart

Which command can I use on a redhat based system to find the dependencies of a package?

finhack

rpm -qpR
Which command will prevent users from exiting a shell script using ctrl c or ctrl z?

W Stewart

the trap command with signals 2 and 20

ChooseLife

Let's revive this thread, shall we?

Question: How can you tell uptime of a system, without issuing "uptime" command?

Expect

ChooseLife wrote: »

Let's revive this thread, shall we?

Question: How can you tell uptime of a system, without issuing "uptime" command?

top -n 1 | head -1

Question: how can you add timestamps to the history command output?

ChooseLife

Expect wrote: »

top -n 1 | head -1

How else? I am not looking for a command that readily prints the uptime (i.e. no "uptime", no "top, no "w", etc)

hiddenknight821

Expect wrote: »

Question: how can you add timestamps to the history command output?

GREAT QUESTION! I kept the honor code, and try not to refer to Google or the like, and I was able to figure out the answer after spending an hour on this. As always, the answer is in the man page of bash since history is bash built-in command. I had to change the HISTTIMEFORMAT variable. Yet, I didn't know how to manipulate it until the man referred me to the third section of strftime man page, which helped me arrive at the answer.

I like the format below. It's more readable:
%D' '%I:%M:%S' '%P' '%Z': '

Expect

ChooseLife wrote: »

How else? I am not looking for a command that readily prints the uptime (i.e. no "uptime", no "top, no "w", etc)

files under /proc are created on boot so their creation date is the same as the uptime of the system
e.g.:

dolev@ubuntu:~$ stat /proc/uptime 
  File: ‘/proc/uptime’
  Size: 0             Blocks: 0          IO Block: 1024   regular empty file
Device: 3h/3d    Inode: 4026532042  Links: 1
Access: (0444/-r--r--r--)  Uid: (    0/    root)   Gid: (    0/    root)
Access: 2015-03-10 06:02:04.864524011 +0200
Modify: 2015-03-10 06:02:04.864524011 +0200
Change: 2015-03-10 06:02:04.864524011 +0200
 Birth: -

Expect

hiddenknight821 wrote: »

GREAT QUESTION! I kept the honor code, and try not to refer to Google or the like, and I was able to figure out the answer after spending an hour on this. As always, the answer is in the man page of bash since history is bash built-in command. I had to change the HISTTIMEFORMAT variable. Yet, I didn't know how to manipulate it until the man referred me to the third section of strftime man page, which helped me arrive at the answer.

I like the format below. It's more readable:
%D' '%I:%M:%S' '%P' '%Z': '

good work,

Question: how can you create an iptables firewall rule related to an already established connection?

ChooseLife

Expect wrote: »

files under /proc are created on boot so their creation date is the same as the uptime of the system
e.g.:

dolev@ubuntu:~$ stat /proc/uptime 
  File: ‘/proc/uptime’
  Size: 0             Blocks: 0          IO Block: 1024   regular empty file
Device: 3h/3d    Inode: 4026532042  Links: 1
Access: (0444/-r--r--r--)  Uid: (    0/    root)   Gid: (    0/    root)
Access: 2015-03-10 06:02:04.864524011 +0200
Modify: 2015-03-10 06:02:04.864524011 +0200
Change: 2015-03-10 06:02:04.864524011 +0200
 Birth: -

I will take it I was looking for something along the lines of

ps eo lstart -p 1 

but this is same

Expect

ChooseLife wrote: »

I will take it I was looking for something along the lines of

ps eo lstart -p 1 

but this is same

Like the ps solution!

Xavor

misposted, delete pls

W Stewart

Haven't been hear in awhile. What does the D status flag in the top command mean?