-4

We don't use logrotate in our legacy servers. So, I want to use logrotate with a cron job. But my team insists on doing something else. i.e. just writing a cron and a random one line script that finds *.log and removes them.

The problem is that I'm unable to tell exactly what are the advantages of using logrotate compared to using that cron one liner. How do I convince my team to use logrotate?

Consider that I am new to this team, about 4 months. I really love the cool technology and I have an intuition that "cool tech" helps in solving most of our tickets. The sad thing is nobody trusts in me that I can write a logrotate configuration although I've studied hard and written a logrotate configuration. I've tested them on local machines as well.

8
  • 41
    "Problem is that I'm unable to tell exactly what are the advantages of using logrotate compared to using that cron one liner" If you, the person proposing that they change to a different technology, doesn't know what the advantages of this new technology is over the currently used technology, then you have no business trying to convince them to change technologies.
    – sf02
    Aug 23, 2023 at 13:48
  • 1
    Yes, I am very passionate about this so I'll implement this next time when I host my own website in homelab. I've an inituition that this will be better but I exactly don't know how. I will ask them for the one liner script. I agree I've no business. I am trying to do the best for company. Everyone knows logrotate is better. Aug 23, 2023 at 13:50
  • 12
    "'cool tech' helps in solving most of our tickets." That's useful or good tech then, usually "cool tech" is what produces the tickets in the first place. If the only reason you're changing a running system is to introduce "coolness" the change probably isn't useful. (Not saying logrotate doesn't offer anything but cool factor here, it has its merit but I can understand if your colleagues don't want to introduce it to old systems just for the sake of it)
    – LW001
    Aug 23, 2023 at 13:52
  • 2
    Possible duplicate of workplace.stackexchange.com/questions/192717/…, at least as far as the answers go. Better, and worth the investment to migrate to right now, are different questions. Put it on the issues list, let it get prioritized with everything else.
    – keshlam
    Aug 23, 2023 at 14:16
  • 9
    To be fair to them, logrotate has only been around for 27 years, so it's understandable that your team are a bit hesitant about deploying such a bleeding edge bit of tech..
    – Gh0stFish
    Aug 23, 2023 at 17:20

5 Answers 5

24

Problem is that I'm unable to tell exactly what are the advantages of using logrotate compared to using that cron one liner.

This is really the fundamental issue. If you're just recommending something because it's "cool" or because other people use it, then you're not going to convince anyone. Especially if you're new to the team, and you have limited experience.

How do I convince my team to use logrotate?

You need to explain to them what problem logrotate solves that they currently have.

Is it simpler to set up? More reliable? Easier to monitor? Does it handle edge cases that the current script fails on? Does it have beneficial features that the current script lacks (like compressing old logs)? Does it help achieve some compliance requirements?

Unless you can do that, no one will be interested.

3
  • 7
    "Especially if you're new to the team, and you have limited experience." And when you admit yourself your manager has lost trust in you and you are applying for roles elsewhere. Fix your fundamentals first before deploying "cool tech" just because you think it is cool. Aug 23, 2023 at 18:05
  • 6
    On top of this, you should be interested yourself in why logrotate is good. Aug 23, 2023 at 19:00
  • 3
    IMO a one liner cron is MUCH easier to understand than any application that requires you to "write a... configuration" Why would I run an app and load a configuration when a one liner does the job?
    – Nelson
    Aug 24, 2023 at 1:02
7

You've overlooked several factors that should bias you against technology changes:

Risk: No tech is bug-free. What your team has at the moment has had most of the bugs knocked out of it the hard way, and the rest they have chosen to live with as acceptable. Switching over to NewCoolTech brings a new set of unknown bugs and maintenance costs.

Inconsistency: Most large companies have bulky, sprawling codebases. Have you got time to migrate everything to NewCoolTech? If not, then you've just added one more thing on to the list of stuff future developers have to learn to be productive.

Opportunity costs: Can you just drop NewCoolTech in with a couple of lines of config change? Unlikely. Regardless, the time spent switching over, and any consequent work coming out of that switch is time that you're not spending fixing bugs and shipping features that actually affect the company's bottom line. You need to earn your salary!

None of this says that NewCoolTech is a bad thing per se, but it should not be your focus unless you can clearly convey what benefits there are that will trump the above concerns.

2
  • 2
    logrotate is ancient tech that was developed specifically because everyone kept writing their own scripts and overlooking some corner case, and is commonly used by Linux distributions for system log files, so my expectation would be that the homebrew solution would have more risk and more inconsistency than the standard tool. Aug 24, 2023 at 5:24
  • In this case, perhaps (although no doubt it still has quirks for the unwary) but as a general rule the point holds. Not sure why OP describes it as "cool tech" then! Aug 24, 2023 at 17:36
4

The problem is that I'm unable to tell exactly what are the advantages of using logrotate compared to using that cron one liner. How do I convince my team to use logrotate?

If you don't know what the advantages are, then how do you expect others to believe in it. I would suggest

  1. Read up (and ask on forums) and make up a chart of pros and cons
  2. Invent a tiny project and implement it both ways (with and without logrotate)
  3. Next time there is a meeting, request a few minutes to demo something
  4. Then demo it
2
  • +1 for a concise step by step approach. I would just insert a step between 2 & 3, which is to think about the difference between logrotate and non-logrotate and decide whether logrotate makes sense.
    – DaveG
    Aug 24, 2023 at 14:01
  • I don't know how do I show it. Aug 24, 2023 at 14:04
3

The way that most groups change technology is that the person in charge decides that it is a good business decision to make the change. For most organizations, that decision does not happen until the organization is facing an existential crisis. In most cases, it is cheaper and better to stay with an existing technology than to switch technologies no matter how much "better" the new technology claims to be.

If you are not the person in charge, then you will need to convince that person of the essential business reason to switch. "Coolness" doesn't work. (I tried and got nowhere.)

This lack of existential business crisis is why you will find businesses still running on old versions of Visual Basic (and perhaps even running on DOS) and old versions of Windows.

How to get a business to look at new technology? Often the best way is to be part of a team developing a new product. New products are often the way that a business will experiment with a new technology and see how well the team can work with it. New products are also the way that businesses react to changes in the marketplace and to new competitive challenges. Unfortunately, rarely do new team members get to be on new product development teams. But you can offer suggestions from outside. (Be prepared for your suggestions to be ignored.)

In other words, you can't do what you want. Use the cool technology on your own side projects and enjoy things there while using the old stuff at work.

2

From my experience, not all the new tech is here to stay, especially lately

I would suggest to do a research and make a little presentation / review of the tech you want to introduce

it may become apparent during your research, that a new tech is not the best fit and doesn`t need to be implemented

1
  • 1
    Also, todays new tech is tomorrows old tech.
    – Abigail
    Aug 25, 2023 at 20:00

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .