Start Minecraft server on RAM disk Linux

Create tmpfs ramdisk.  Note if your Linux user is something other than steve you’ll need to change where appropriate.

mkdir /home/steve/mcdisk

In etc/fstab add the following

tmpfs /home/steve/mcdisk tmpfs defaults,size=4096m 0 0

This creates a 4GB ram disk at /home/steve/mcdisk

To mount it you can either reboot, or run

mount -a

Copy your current Minecraft directory to the ram disk

cp -R /home/steve/Current_MC_Server/ /home/steve/mcdisk

Create a Bash script in “/home/steve” named “ramdisk_save.sh”

Paste the following in.  You may need to install rsync if you do not have it installed

!/bin/bash

RAMDISK="/home/steve/mcram/"
MCDIR="/home/steve/1.13"

rsync -r -t $RAMDISK/ $MCDIR/
rsync -r -t $MCSTORE/ $MCPATH/

Now add the script to crontab

crontab -e

and

 */5 * * * * /home/steve/ramdisk_save.sh

This will now run every 5 minutes and sync any changes on the ram disk to the original directory.

Start the Minecraft server

java -Xmx3072M -Xms3072M -jar server.jar nogui

Ubuntu UniFi server running out of space on /run

Had an issue that /run was randomly running out of space which in turn would interfere with the unifi-video service causing it to run, but not record.

/run looks like a tmpfs or ramdisk that Ubuntu sets up.  So you can do a “temporary” fix by remounting the tmpfs with a larger size.  Example below.  If /run is a 2GB directory, you can remount changing the size from 2GB to 2.5GB.

sudo mount -t tmpsfs tmpfs /run -o remount,size=2500M

Note that it is a temporary fix and goes away after a reboot.

The issue ended up being that the WiFi UniFi controller was setup to auto backup everything once a week.  So as it was backing stuff up, it would eat up the available space in the tmpfs, think there may be an issue with the size of the UniFi data and maybe not being able to fit it all in RAM?

Running the following command

df -h --max=1 /var | sort

shows the following

1.1M /run/udev
2.5G /run/
2.5G /run/unifi   <-- UniFi controller
4.0K /run/initramfs
8.0K /run/network
12K /run/user
288K /run/samba
404K /run/systemd

Looking inside the unifi directory shows the following folders.  Looks like the they are temp files.

200M /run/unifi/ExpTmp351719567129045774
696M /run/unifi/ExpTmp3406220793759111216
1.6G /run/unifi/ExpTmp3368400690321364109
0 /run/unifi/work
2.5G /run/unifi

Running an ls inside the folder shows

-rw-r----- 1 unifi unifi 13971807 Jul 2 02:30 db.gz
-rw-r----- 1 unifi unifi 1665223462 Jul 2 02:56 db_stat.gz

Looking inside the UniFi controller it is set to auto backup on Monday at 2:30AM

Looks like “/run/unifi” is used as a temporary folder to create the backups and when it is completed converts it to a .unf file and moves it to “/usr/lib/unifi/data/backup/autobackup/”  So the left over temp files were never completed or something caused them to stop working.  Maybe the backup was big enough to take up all the tmpfs space and cause the backup to fail.

Resolution.  To fix the problem I turned the data retention down to a week and deleted the temp files in /var/unifi.  Had to sudo su to root.  sudo wasn’t cutting it for some reason.

sudo su
rm -rf /var/unifi/Exp*
exit

Should be all good now.

DNF/YUM not working in chroot environment

The reason is probably because the chrooted environment can’t resolve DNS.

Test it with

ping incredigeek.com

If it is not resolving, edit “/etc/resolv.conf” and change/add your nameserver.  Or just replace everything in it with

echo "nameserver 4.2.2.2" > /etc/resolv.conf

It should now be able to resolve and you should be able to use yum, or dnf.

yum update

RAID Volume not accessible in Linux

Typically your RAID volumes will show up as /dev/mdXXX

If it is not, it could be because the device mapper module is not loaded.  Load it by running the following command.

modprobe dm-mod

As a side note you can list the block devices using

dmraid -b

https://bbs.archlinux.org/viewtopic.php?id=42321

If your still having trouble you can try installing mdadm and dmraid

apt-get install -y dmraid mdadm

 

Delete SNMPv3 User on Linux

Don’t know if this is the recommended way to delete a user, but it seems to work.

sudo service snmpd stop

Open up the snmpd.conf file in /var/lib and find the line with the SNMP user and delete the line

sudo vi /var/lib/snmp/snmpd.conf

Save, exit, and start snmpd

sudo service snmpd start

These steps work for Ubuntu, but should work for any Debain based distro as well as CentOS, Fedora, RedHat etc.