[OpenIndiana-discuss] So is Solaris officially dead?

Alexander Pyhalov alp at rsu.ru
Thu Sep 7 05:48:22 UTC 2017


Hi, people.

We all liked Solaris, but its fate to become new HP/UX was evident from 
day 1, when Oracle cancelled OpenSolaris.
There's nothing to discuss here, it was evident that it'll become some 
OS which nobody has never seen sooner or later
(just try to search for Solaris administrator job on LinkedIn and 
compare to Linux administrator job offers).

But this list is not about Solaris.
Let's look at consequences.
1) We are loosing the great source of upstream fixes for a broad range 
of software. On itself it's not catastrophic, but in areas where it was 
the only source, it is
(read, desktop illumos will die rather sooner then later). Fortunately, 
we were successful in incorporating our changes in Mate and 
Enlightenment, but it's not full
story (for example, try to build and use polkit on illumos...). I hoped 
there were enough interested parties to support server software on 
illumos/Solaris systems,
but there are issues even in this area. We can look at MongoDB decision 
here, which was not surprising. Now suppose if PostgreSQL devs do the 
same. So, we need upstreaming
our patches everywhere where it's possible. We (devs and users) need to 
communicate to upstream projects, so that they don't forget about our 
existence.
2) Another underestimated issue is lack of documentation. Our (or 
SmartOS, or OmniOS) documentation is nothing compared to the Oracle 
Solaris documentation library. Sometimes
it's very unevident how to configure some subsystems (for example, 
configure FMA to send events about failed services via email). We (whole 
illumos community) have no documentation team, and this is an issue. 
Compare OI 'Handbook' to FreeBSD one to see the difference. We have 
infrastructure for creating documentation 
(https://github.com/OpenIndiana/oi-docs/), but it has just about 20 
commits for the whole 2017 year. And most of this documentation discuss 
beginners topics, which are not of great interest to system 
administrator, who tries to solve exact problem. This leads people to 
look at code as on only exact source of documentation, and this is 
unacceptable. This is the area where usual users could help.
3) We all speak that Oracle didn't support Solaris in Universities, but 
what did illumos community to support illumos distributions there? For 
example, FreeBSD has repository https://github.com/teachbsd/course with 
materials for Cambridge FreeBSD course. Do we have a single man who 
tried to teach OS basics on base of some illumos distribution or to make 
it easier for others to do?

Another (not directly related issues):
4) We efficiently have no translations to other languages (where we have 
translations, they come from upstream projects).
5) After OmniTI exiting illumos family, we efficiently have no entity, 
providing commercial support for general purpose solarish OS.

To Adam:
6) Please, don't forget to create our mirror of Solaris-related Oracle 
github repositories ;)

Perhaps, instead of discussing sad but expected event, we could start 
solving our own problems?
---
System Administrator of Southern Federal University Computer Center




More information about the openindiana-discuss mailing list