On a VPS, run the same CPU benchmark at three different hours and watch st in top. If scores swing while st rises, shared CPU time is a likely cause. Rule out cron jobs and disk too
Avant VDS-2 in Flint: 2 dedicated cores, 8 GB RAM, $40/mo😉
7
1/ Four checks to run at home before you ship a server to a colocation facility. Each is easier to fix on your desk than after the box is racked😉
👇
1
4
0 9 * * * runs at 09:00 server time.
Check: timedatectl
Set: sudo timedatectl set-timezone UTC
At DST changes, a job in the shifted hour can run twice or skip. UTC has no DST.
Avant VPS, full root, so you set the time zone. VPS from $5/mo in Flint😉
10
1/ A Linux server can hit its RAM limit during a build, a backup, or a traffic spike. Then the OOM killer ends a process, usually the one using the most memory, often a database
A swapfile gives the kernel some slack first. Setup in the next post👇
1
20
GTA VI preload math: Sensor Tower estimates 12M+ preorders through Nov 13 (via CNBC), and Forbes lists up to 212 GB of space per preload
If every preorder pulled it once, that's about 2.5 exabytes😵💫
33
"No space left on device" while df -h shows free space usually means the filesystem ran out of inodes.
Check with df -i. IUse% at 100% is the problem. On ext4 the inode count is set when the filesystem is created, so the usual fix is fewer files
👇
1
9
The usual cause is millions of tiny files: PHP sessions, mail queue entries, cache directories, or one log file per event
To see which directories hold them:
sudo find / -xdev -type f | cut -d/ -f1-4 | sort -T /dev/shm | uniq -c | sort -nr | head
👇
1
7
