EX200 interview questions & answers

Permissions are read/write/execute for owner, group and others. For a shared directory, set group ownership (chgrp), give the group rwx, and set the setgid bit (chmod 2775) so new files inherit the group. A sticky bit (chmod +t) on a shared directory stops users deleting each other's files.

SELinux context. Apache can only serve files with the right type (httpd_sys_content_t). Check with ls -Z and the audit log (ausearch/sealert), then fix with restorecon or set the correct context with semanage fcontext. Never just disable SELinux — set the context or the right boolean.

Extend the volume group with a new physical volume if needed (pvcreate, vgextend), grow the logical volume (lvextend -r, which also resizes the filesystem), and verify with df. LVM lets you grow online without repartitioning, which is why production systems use it.

Read the error — usually a bad /etc/fstab entry or a missing device. Remount root read-write (mount -o remount,rw /), fix or comment the offending fstab line, and reboot. If it is a root-password reset you boot to rd.break, chroot, passwd and touch /.autorelabel. systemctl status and journalctl -xb show what failed.

systemctl enable --now <service> both starts it and enables it for boot. Confirm with systemctl status and journalctl -u <service>. enable alone only sets boot behaviour; start alone does not survive a reboot — people mix these up constantly.

top/ps to find the PID and user; renice to lower its priority or kill/systemctl stop if it is a runaway. For repeat offenders, cap it with a systemd slice or cgroup limits. Check logs to find why it spun before killing it blindly.

With nmcli: nmcli con mod <conn> ipv4.addresses, ipv4.gateway, ipv4.dns, set ipv4.method manual, then nmcli con up. NetworkManager keyfiles persist it across reboots. Editing ifcfg files by hand is legacy and error-prone on current RHEL.

cron is simple and familiar for time-based jobs. systemd timers add dependencies, accurate logging in the journal, missed-run catch-up (Persistent=true), and randomised delays — better for servers and for jobs that must run after boot or after another unit. On RHEL, timers are the modern choice.

sudo with a least-privilege entry in /etc/sudoers.d/ — allow only the specific commands they need, and log usage. Add them to a group and grant the group rather than editing individuals. This is auditable and revocable, unlike sharing the root password.

Most likely inodes are exhausted (df -i) from millions of tiny files, or a deleted-but-still-open file is holding space (lsof | grep deleted) — the space frees only when the process is restarted. Occasionally it is a separate full filesystem mounted under the path. Check inodes and open deleted files first.