So the server I have hosted at Dacentec is starting to show its age and limitations. Luckily I can get my hands on some decent hardware thanks to ebay and other methods. So the past few days when time allows I have been working on building a replacement 1u server to what I currently have. In doing so I am taking the RAM from 16GB to 196GB. With that said the system is installed with a stock ubuntu 17.10 server using guided lvm partitioning. Which thinks like days of old where total ram equals the swap size needed.
Avoiding the whole debate of how much swap would be ideal. I am just going to scale it back to about 16GB. Thus freeing up about 80% of the OS RAID 1 drives.
As always google and a stackoverflow/askubuntu (Ref links below) along with some good old fashion thinking.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 root@atlas:~# lvs LV VG Attr LSize Pool Origin Data% Meta% Move Log Cpy%Sync Convert root atlas-vg -wi-ao---- 86.88g swap_1 atlas-vg -wi-ao---- 191.95g root@atlas:~# swapoff /dev/atlas-vg/swap_1 root@atlas:~# lvreduce --size 16G /dev/atlas-vg/swap_1 WARNING: Reducing active logical volume to 16.00 GiB. THIS MAY DESTROY YOUR DATA (filesystem etc.) Do you really want to reduce atlas-vg/swap_1? [y/n]: y Size of logical volume atlas-vg/swap_1 changed from 191.95 GiB (49140 extents) to 16.00 GiB (4096 extents). Logical volume atlas-vg/swap_1 successfully resized. root@atlas:~# mkswap /dev/atlas-vg/swap_1 mkswap: /dev/atlas-vg/swap_1: warning: wiping old swap signature. Setting up swapspace version 1, size = 16 GiB (17179865088 bytes) no label, UUID=f0260f73-bf13-41d4-af79-3203114b3f9d root@atlas:~# swapon /dev/atlas-vg/swap_1 root@atlas:~# lvs LV VG Attr LSize Pool Origin Data% Meta% Move Log Cpy%Sync Convert root atlas-vg -wi-ao---- 86.88g swap_1 atlas-vg -wi-ao---- 16.00g