🚗🏍️ Welcome to Motoshare!

Turning Idle Vehicles into Shared Rides & New Earnings.
Why let your bike or car sit idle when it can earn for you and move someone else forward?

From Idle to Income. From Parked to Purpose.
Earn by Sharing, Ride by Renting.
Where Owners Earn, Riders Move.
Owners Earn. Riders Move. Motoshare Connects.

With Motoshare, every parked vehicle finds a purpose. Partners earn. Renters ride. Everyone wins.

Start Your Journey with Motoshare

Three Ways to Run a Client’s Linux Servers, Compared for Freelancers

You land a small client whose three Linux servers have been looked after by “whoever was around”. Now they are yours. Before you bill a single hour you need a way of working, and SysOps, the day-to-day care of individual hosts, gives you three realistic options. This piece compares them without hype, for someone who is paid per engagement and cannot afford surprises in the middle of the night.

The ground rules a freelancer should settle first

Whatever route you choose, agree on three things in writing: who owns the access keys, what you are allowed to change without asking, and how a change gets recorded. The Ops Atlas lists unrecorded hand edits and shared administrator logins among its anti-patterns, and for a freelancer both are also commercial risks. If something breaks and you cannot show what you touched, the conversation about fault gets awkward fast. If you want a broader view of keeping client relationships steady, a practical article on winning confidence with a new customer covers the people side.

Option one: log in over SSH and make the change by hand

What it is: connect, edit, restart, disconnect. Everyone starts here.

In its favour: zero setup, zero tooling cost, and it is fast for a one-off fix on a single machine.

Against it: nothing is repeatable and nothing is recorded unless you are disciplined. Two servers that were supposed to match will drift apart in a month. There is also a safety angle. The Atlas notes that for the remote-access daemon, the first value set for most keywords wins, so an edit appended at the bottom of a file may be ignored without any error. Its advice is to run the daemon’s built-in syntax test and to leave a spare login session open until a fresh connection has succeeded. Hand work is workable only if you treat that checklist as non-negotiable.

Verdict: acceptable for emergencies and for a single host, poor as a default.

Option two: describe the baseline in Ansible and re-apply it

What it is: write down the intended state of the machines as playbooks and let a tool enforce it.

In its favour: idempotent tooling can safely be run on a schedule, so the Atlas points out you can re-apply the baseline every hour without fear. Your work becomes a repository you can hand to the client, which doubles as documentation and as a deliverable they can see.

Against it: there is an upfront cost in learning and in writing the first version, and a client with two tiny servers may not want to pay for it. You also need to be strict about not editing machines by hand afterwards, otherwise the playbook and reality part ways.

Verdict: the strongest long-term value for a client with more than a couple of machines and an ongoing retainer. It is also the easiest option to hand over when you move on.

Option three: run commands through a cloud provider’s remote-command service

What it is: with machines on a public cloud, use the provider’s run-command feature and session manager instead of opening inbound ports.

In its favour: no keys scattered across laptops, central logging of what ran, and no need for an open management port. For a freelancer who works from several networks this removes a lot of friction.

Against it: the Atlas warns that the session tool’s default is a shared account with passwordless administrator rights. Unless you configure the “run as” tags and tightly scoped permissions, you have traded one weak pattern for another and every session looks the same in the audit trail. It also only works where the provider’s agent is installed and reachable.

Verdict: strong when configured properly, weak if left on defaults. Budget time for the permission setup, and note that the control plane itself belongs to a separate discipline, so check what your client’s cloud engineer already manages.

Side-by-side summary

  • Speed on day one: hand edits win, remote-command second, playbooks last.
  • Repeatability: playbooks win by a distance.
  • Audit trail: playbooks in version control and a properly scoped remote-command service both beat memory and shell history.
  • Handover to a client: playbooks are easiest to pass on; hand work is hardest.
  • Cost to start: hand edits are free, the other two cost setup time.

What it costs the client

Freelancers often skip the money conversation, yet it shapes which option is sensible. Hand edits look cheap because the hours are small, but the client pays later in outages that nobody can trace. Playbooks cost more in the first invoice and less in every invoice after it, since a repeat task becomes a re-run instead of a rediscovery. Remote-command tooling sits in between: modest setup, then a quiet per-task saving. When you quote, show the client the shape of the curve, not only the first number, and offer the option that matches how long the engagement will really last.

The things that should not depend on your choice

A few practices apply whichever route you take. Keep SELinux or AppArmor enforcing and relax one domain or profile if something needs room, rather than switching protection off for the whole machine. Prefer systemd timers over cron for client jobs, because they can retry missed runs and trigger a failure hook. Check disk runway weekly with a simple calculation so you can tell the client “you have about a week” before it becomes a crisis. And rehearse a restore at least once, since the Atlas is blunt that a snapshot is not a backup.

Reading about someone else’s setup is also useful perspective for your own pricing and scope. Freelancers weighing up how much to take on alongside other work may find the side-hustle guide for people who keep a day job relevant, because maintenance retainers are exactly the kind of work that can swallow evenings if nobody sets limits.

How I would decide

For a single host and a one-off job, do it by hand with the safety checks. For two or more hosts that you will touch repeatedly, invest in playbooks. For cloud machines where inbound access is a concern, use the provider’s remote-command service and fix the permission defaults first. Where one client uses all three, resist the temptation to pick per mood; write down which route governs which machine. The Atlas entry covering SysOps practice has the underlying checklists if you want to build your own handover document from them.

The common thread is boring but valuable: choose the option that leaves the clearest record, because a clear record is what lets you sleep and what lets the client trust you with the next machine.

Related Posts

Eight Cluster Chores Every DevOps Freelancer Can Offer Clients

Eight small, billable Kubernetes checks a freelancer can bring to a client: dates, requests, budgets, restores and a thirty-sixty-ninety plan.

Read More

InfraOps Problems Freelancers Inherit, and How to Fix Them

Five common symptoms a freelance engineer meets in a client’s infrastructure, each paired with a cause and a remedy, plus the evidence worth recording.

Read More

Six CloudOps Misunderstandings to Sort Out Before a Client Call

Certificates, renamed roles, two-provider setups and uptime promises: six CloudOps misunderstandings worth untangling before you scope a freelance contract.

Read More

Turning DisasterRecoveryOps Into a Freelance Specialty

How independent engineers can learn, package and price disaster recovery work, using a maturity ladder and a ninety-day offer.

Read More

BackupOps FAQ for DevOps Freelancers: Questions to Settle Early

Eleven straight questions and answers on restore responsibility, credentials, scoping and quoting for independent DevOps contractors.

Read More

A Freelancer’s Nine Steps to Tackle ObservabilityOps for a Client

A story-led, step-by-step method for freelancers who need to fix a small client’s alerts, logs, traces and monitoring bills in one short contract.

Read More
0 0 votes
Article Rating
Subscribe
Notify of
guest
0 Comments
Oldest
Newest Most Voted
0
Would love your thoughts, please comment.x
()
x