v7.1rc1 [development] is released!
Mon Aug 23, 2021 1:13 pm
We believe this “release candidate” version is pretty much ready for “stable” release on all devices. If you experience any issue, please open a support ticket with as much details as possible and supout.rif file attached. Faster we can acknowledge all active issues, faster we can resolve them and release a “stable” version.
Would it be unwise to do it?
I can image that many things around ROS need to be working and that demands time getting stable.
But given my use-case, is it likely to run into stability issues?
It is completely inadequate, and devoid of any logic, to use “rc” software in production.
But it’s even worse to use a version (any) that just came out two days ago for something to put into production.
Really absurd.
Only worth moving to v7 if there’s a feature you absolutely NEED for some reason - even then its a silly idea.
I’ll be on 6.x LT releases for probably quite a while (like 7.10 stable) - as also with the fundamental changes in a lot of the logic and objects, it’ll be a ‘whole network migration’ plan to move to ROSv7 I feel.
CRS3xx will have L3 HW offloading in ROS v7 (currently in release candidate stage) where inter-VLAN routing will be offloaded to hardware switch chip and it will be possible to have wire-speed connections between devices in different VLANs/IP subnets.
Also, since I’m designing a new network, starting with ROS7 instead of first ROS6 and then upgrading to 7 (when it soon(?) becomes stable) saves time & effort.
Since @emils is posting (see top of my OP):
We believe this “release candidate” version is pretty much ready for “stable” release on all devices. …
(Also, it seems weird to me that ‘RB5009UG+S+IN’ will be available very soon and is ROS7 only; does that mean new hardware out to the market with testing (unreliable) software for production…?)
Anyway, based on the input so far: others take the comment of @emils with a grain of salt and advice me to save me pain & trouble (in addition of time & effort).
@BartoszP: I was referring to your comment regarding
Is it really so hard to read this topic…
Since I am quoting from that exact topic, it is reasonable to assume I did already read the topic you are referring to.
You have to judge for yourself if RC version is equal to “production ready”.
I am asking for input on the experience of others, in relation to the post of @emils regarding:
We believe this “release candidate” version is pretty much ready for “stable” release on all devices.
So for whom & why Mikrotik supports “stable” and “long term” versions in 6.x line?
There might be several reasons: one of them is for custormers who need patches on there devices, but are not willing to adapt configurations and scripts to a new ROS version. Running a Debian ‘testing’ release doesn’t make an admin a fool by default for example. I don’t know what to expect from ROS, that’s why I am actually asking others with more experience on ROS.
On Aug, 25 you refered to the post from 23rd of August and it’s resonable to assume that you had not read tens of posts sharing information: “problem…”, “not working…”, “missing …”
If you ask if it would be resonable to put ROS 7rc1 into production it’s obvoius to me that you like “livin’ on the edge”.
This sort of broad sweeping statement is not accurate and appropriate in all cases. There's been many 'release ready' RouterOS versions that have had glaring issues and bugs
Just because something is labelled 'bugfix' (or whatever the current term is) doesn't mean stable and production ready. It's still a case of 'we think its stable, probably'
The only sensible sweeping logic is to first ask
what's the current situation, is everything stable or is there already a problem that needs addressing?
what features, fixes or improvements are in the new version?
what is the service impact/risk?
Those 3 things will change the strategy. In some environments yes absolutely the very latest RC should be applied immediately. If the network is already on fire the risk of anything worse is greatly outweighed by the potential to fix an issue. And if there's a desperately needed feature it again may be worth doing
Conversely if there's nothing inherently vital in updates, everything is perfectly stable and its a highly important segment then you probably shouldn't update AT ALL, especially with MikroTik
The only solid answer is 'it depends'. Circumstances always trump broad sweeping 1-size-fits-all statements
