Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sun, 21 Dec 2014 22:57:18 +0300
From: gremlin@...mlin.ru
To: owl-users@...ts.openwall.com
Subject: Re: owl-startup

On 2014-12-21 07:39:10 +0300, (GalaxyMaster) wrote:

 >> While working with one project, I faced the inconvenience caused
 >> by ancient init scripts in owl-startup (once again). However,
 >> this time I had several hours to perform some rewrite...
 > What would help a lot (and possibly create more traction) is
 > providing some rationale behind the rewrite.

1. All possible network configurations
2. Support for loopback/cryptoloop (including trivial `umount -a -rdf`)

Of course, some other minor issues were fixed by the way.

 > The way you did it is like "here is a cat in a box: you may try
 > to play with it but your mileage may vary".

Yes, as it's exactly that. And my FTP server logs show that people
are really interested in such experimental packages.

 > After such a proposal, personally, I don't have any desire to
 > try my luck.

You aren't requred to do that: although your feedback could be
really valuable, for now I'd be quite happy with functionality
tests - that's why I've posted this to -users@

 > On a side note, I would never imagine that I'd raise it here
 > and be a proponent of it, but here it comes: should we start
 > working toward integrating systemd into Owl?

Not earlier than at least the following packages would be updated
|| added:
* kernel
* openssh
* pkgconfig
* nginx
* httpd (with full LAMP stack)
* rsync
* qemu

In general, these suggestions should go to -dev@, but once you've
started this discussion in -users@ - well, it would be fine to let
our users know what we're working on.

Several more words on Qemu: as Owl is really minimalistic system,
it just perfectly suits the position of virtualization host. That
means, we could move in this direction and be in-trend.

 > Over the last 6 months I was kind of "forced" to work extensively
 > with distros that switched to systemd. To do my job properly I
 > had to learn the design of that framework and it really looks
 > logical and once you jump through the hoops of the learning curve
 > you cannot deny that Poettering and Co did a huge amount of work
 > to standardise the startup & init process. The documentation is
 > also _very_ good.

Personally I see only one application for systemd: when you make
the system a bloatware, it starts up very slowly, and the systemd
attempts to speed it up.

Owl with its' two hundred packages starts in 10...15 seconds, so
it's hardly a bloatware and doesn't need the systemd.

 > Sooner or later we will face the fact that we should decide
 > whether we go with the trend or make our own path. However,
 > knowing our limited resources I doubt that we are going very
 > successful with the latter. Opinions?

Actually, we've already made our choise: keep small, secure and
effective, using OpenVZ as a task separation technique (instead
of selinux etc). Just adding Qemu/KVM support (which I'm working
on) will make it the complete virtualization solution, supporting
both VPSes and VDSes.


-- 
Alexey V. Vissarionov aka Gremlin from Kremlin <gremlin ПРИ gremlin ТЧК ru>
GPG: 8832FE9FA791F7968AC96E4E909DAC45EF3B1FA8 @ hkp://keys.gnupg.net

Content of type "application/pgp-signature" skipped

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.