mintupgrade check won't take yes for an answer

[jcoles]

I'm trying the process to update from 17.3 to 18. After 20 minutes or so, "mintupdate check" produces no warnings that I can see and asks "Do you want to continue?[Y/n]. I answer "y" and hit return. The terminal shows me that I said "N"! It then proclaims "Abort" and "Restoring your backed up APT sources". What's wrong? Why does it ask me this question and then ignore my answer?

[Cosmo.]

Perhaps this is a misunderstanding. mintupgrade check is a simulation for the upgrade, it does not execute the upgrade. So if the check has been completed (seems so after 20 minutes), the job is done and the previous repos get restored. So far this is as intended.

Only if you continue with the next commands (mintupgrade download and mintupgrade upgrade) the sources will be permanently changed and the upgrade gets executed.

[trytip]

if mint is not bleeding edge software, why are so many rushing into upgrading to mint18? if it ain't broke why break it? maybe you should install mint18 fresh instead of upgrading, which is not even supported officially yet

[Cosmo.]

Upgrading is officially supported since 3 days. mintupgrade is a tool by the Mint developers.

[karlchen]

Let me add: Read and follow Clem's tutorial "How to upgrade to Linux Mint 18" carefully.

[jcoles]

Yes, Cosmo, I think I was confused. The "proceed?" question was not about upgrading, but about undoing the simulated upgrade. A check function should end with a clear statement of OK or a list of issues to be resolved. I saw neither. Eventually I just went ahead with the next steps and was successful.

The laptop I tried this upgrade on was not mission-critical. I could afford to mess it up and reinstall from scratch if necessary. I must say that I find the team's cautious "do you really need to upgrade?" approach suggests a lack of confidence. Don't they want us to try their latest version? Having seen the upgrade announced in the blog, I gave it a try. In retrospect, I wouldn't bother with the upgrade process again. It takes several times longer than a fresh install. My only beef so far is that Mint 18 takes over 1.5 minutes to start, about twice the time that 17.3 required. On my desktop, there was no increase in startup time after upgrading to Mint 18 from a clean install. I'll investigate some more. I can try a clean install, or even re-install 17.3 if necessary.

Thanks, guys.

[Cosmo.]

I must say that I find the team's cautious "do you really need to upgrade?" approach suggests a lack of confidence. Don't they want us to try their latest version?

Every update can imply regressions, the possibility in case of an upgrade is even greater, an upgrade to a newer major version implies the greatest possible risk.
As we have seen in the last weeks, there are systems, which run well with LM 18 but there are others, which do not. Also we can see, that some users complain about bugs, which are in LM 18, others do obviously not see them (probably because they do not use the affected features / programs?). This is all not predictable, so a user must find out for his / her system, if the new version does work as expected or not. Simple fact is, that the new version does not work as well as LM 17.3 on some systems (for example here).

I wouldn't bother with the upgrade process again. It takes several times longer than a fresh install.

In the past there has been several discussion, if the in-place upgrade to LM 18 should be done or not. I always stated, that this is not discussable, until the upgrade method cannot be tested. So I was fairly open about that. I have spend a lot of time for testing the upgrade process and repeated it to ensure, that the results are reproducible. After that I support what you wrote. I do not recommend it, out of several reasons:

1. The needed time for upgrading - without the needed time for downloading - is more than 5 times the time for a fresh installation. But that is not all. If you do a fresh installation, you can leave the computer, after you have covered all requests at the start, and do something else (take a meal, talk with other people, what else). But during the upgrade process you have to answer more than a dozen questions, otherwise the upgrade process will pause. So in the sum the effort of needed time is far more than what is measurable. I also did not measure the time for checking the result of the mintupgrade check, which is officially noted as an "extremely important" step. So at the end you need likely about ten times the amount of time.

2. The result is not identical. " examples: Mint-Y does not get installed via upgrade, samba does not get removed. Most likely there are more differences.

My only beef so far is that Mint 18 takes over 1.5 minutes to start, about twice the time that 17.3 required. On my desktop, there was no increase in startup time after upgrading to Mint 18 from a clean install.

The increased boot time and login time on some systems is indeed a problem with LM 18 / Ubuntu 16.04. Also here we can see very different results on different systems. On my test system (virtualized) I get already with a fresh installation far bigger times needed as well for booting and for login in in comparison to LM 17.3. If I measure and compare this on the upgraded system I see, that the needed time for booting is more than 80 seconds(!). On a fresh installed system this is about the half(!). During the beta testing phase I found out, that on this system samba increases the boot time for about 10 seconds. So if I reduce the 80 seconds by this 10, there is still a difference of 30 seconds in the boot time between a fresh installed 18 and an upgraded 18.

At the end I come to the result, that a fresh installation is the better approach. In difference to the upgrades between point-releases you have in any case to restore PPAs afterwards manually. User settings can get backed up and restored, this is a matter of a few minutes.

Final note: My tests have been done using an LM 17.3 system, which has not been altered from installation. That means, no PPA, no additional software, just plain Rosa; only level 1 to 3 updates have been applied. So I cannot judge with the results, if and what problems may arise on a system, that has been altered by the user before he does the upgrade. But surely it is predicable, that the possibility of problems will not get smaller in those "living" cases.

[jimallyn]

I must say that I find the team's cautious "do you really need to upgrade?" approach suggests a lack of confidence.

Linux users tend to believe in "if it ain't broke, don't fix it."