nf_conntrack table full, dropping packet

Virtualmin
1

Hello,

just got a message on the console:

“nf_conntrack table full, dropping packet” and I see that HDD is almost full, but which folder
is “full” in this case ? the log perhaps or ??

I need to delete these files but I am not sure which folder and location is used in this case…

I am talking about “Local” disc space…

and btw, the Virtual memory is very “low”

Any idea ?

Thank You

Amel

2

Howdy,

It sounds like you have some resource problems there… are you by chance using OpenVZ?

If so, can you post the contents of you /proc/user_beancounters file?

-Eric

3

Hello Eric,

thank You for reply.
It seems that I we don`t have the file named user_beancounters in that location…

And we are not using the OpenVZ…

but the HDD was “filled up” in one day only which is so strange, can`t find out which folder …

Amel

4

You can use the following command to see sums of subdir sizes of the current directory. Start in / and work your way down.

du -sh *

But I think you have multiple problems,considering the nf _conntrack warning and the low memory.

5

thank you for reply, not sure what causes this issue, never seen it before.

The issue is that server have 11.80 GB total, 1.31 MB used RAM mem, so it seems that
for some reason the server is using the local HDD instead of RAM mem ??

Any idea where to look for this issue ?

can You write the full command (du -sh *) ?

Thank you again for reply !

6

This is how it looks like just now…

Virtual Memory:

11.80 GB total, 1.31 MB used

Real Memory:

9.66 GB total, 1.68 GB used

Local Disk Space:

86.81 GB total, 77.94 GB used

7

I see that there is huge amount of data in the var/log, can I delete just everything in
var/log so it can begin from the start again ? so I can follow it…

there is some folders in var/log also for example:

clamav
mailman
proftpd
virtualmin

etc…

can I delete just everything in var/log ?

thank You

8

You’re mixing up “Virtual memory” and “Physical memory”. What you interpreted as “11.8 GB physical memory, and only 1.3 MB being used” is in fact the virtual memory i.e. swapfile.

You memory values look okay; you have 9.66 GB physical memory with only 1.68 GB in use, and swapfile is basically not in use at all.

To make sure, you can use the command free in a shell to get further details. When you post command line outputs here, please enclose them in [code][/code] tags.

The disk usage command is exactly what I wrote:

du -sh *

This shows the usage of all subdirectories, summed up per directory, of your current working directory.

Basically you could delete everything in /var/log, but that is highly discouraged. Several processes will either keep files locked or go crazy when you simply delete their log files. Better find out which files are so large in /var/log:

du -sh /var/log/* ls -lh /var/log

9

sorry, my mistake of course I mixed it.

You have right about log file delete, it`s not recommended at all, just tested it on another test server.
Will try to run the commands you mentioned and try to find out which logs eats the hdd space…

thank you for helping and info !