|
Home > Archive > Unix administration > July 2006 > AIX vs Solaris
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
|
|
| racman 2006-07-20, 1:23 pm |
| I need to give a presentation on the benefits of AIX vs Solaris.
Can everyone share your thoughts. Keep in mind that I`m not looking
to dismiss either platform, I think they both work very well with
specific databases,
storage and apps. Thoughts please!
Thanks
| |
| Dave Hinz 2006-07-20, 1:23 pm |
| On 20 Jul 2006 08:44:21 -0700, racman <arubin@mfs.com> wrote:
> I need to give a presentation on the benefits of AIX vs Solaris.
> Can everyone share your thoughts. Keep in mind that I`m not looking
> to dismiss either platform, I think they both work very well with
> specific databases,
> storage and apps. Thoughts please!
Well, like any other "which OS should I use for this situation"
question, it really comes down to - what apps do you need to run, where
do they run best, and what hardware is available?
Solaris admins are easier to find, at least around here (midwest USA).
AIX has a great admin tool that lets sysadmins do some pretty
sophisticated stuff without needing to be intimately familiar with the
inner workings. (not saying it's wise, just that it's possible). The
answer of "it depends" is a weak one, but without some clarifying
details I think it's the best I can give. And, harder to measure things
like staff comfort level are real but fuzzy.
Dave Hinz
| |
| Doug Freyburger 2006-07-20, 7:24 pm |
| racman wrote:
>
> I need to give a presentation on the benefits of AIX vs Solaris.
Typical uptime is higher on AIX, number of apps available is
higher on Solaris. I think these two have a direct inverse
correlation.
AIX was an early adopter of several important technologies -
logical volumes on the installation media, dynamic kernel
parameters and so on. Solaris now has all of these but the
next cool technology is likely to have AIX as the early
adopter.
AIX LVM is simpler and therefore easier to script. Solaris
has two LVMs but the Veritas one is more complex and
therefore able to do more stuff.
Hardwarewise IBM has spent more effort over the years on
their high end boxes and Sun on their low end boxes so
even though there's plenty of overlap in the product lines you
would likely want Suns for rackfulls of small clustered servers
and IBM for the big database boomers.
Do you know about Bruce's UNIX Rosetta stone? It's a table
of specialized command equivalencies. What it boils down to
is if you can do it on one UNIX-like box chances are you
can on the other types as well. Which is going to be better
depends very much on the exact situation.
| |
| Mogens V. 2006-07-20, 7:24 pm |
| Doug Freyburger wrote:
> Do you know about Bruce's UNIX Rosetta stone? It's a table
> of specialized command equivalencies. What it boils down to
> is if you can do it on one UNIX-like box chances are you
> can on the other types as well. Which is going to be better
> depends very much on the exact situation.
It's been updated quite resently, hasn't it? Last time I looked, it
seemed a bit out of date, IIRC. I may easily be mistaken...
BTW, as part of the upcoming BSD certifications, they have a nice BSD
commands reference, outlining commands similarities/differences between
the four major BSD's:
http://www.bsdcertification.org/dow...rence_en-en.pdf
--
Kind regards,
Mogens V.
| |
| Casper H.S. Dik 2006-07-21, 1:20 pm |
| "Doug Freyburger" <dfreybur@yahoo.com> writes:
>AIX was an early adopter of several important technologies -
>logical volumes on the installation media, dynamic kernel
>parameters and so on. Solaris now has all of these but the
>next cool technology is likely to have AIX as the early
>adopter.
Really? Which cool technology was adopted in AIX recently?
>AIX LVM is simpler and therefore easier to script. Solaris
>has two LVMs but the Veritas one is more complex and
>therefore able to do more stuff.
You're forgetting about ZFS which is much easier than anything.
(It's LVM and filesystem rolled into one)
(One or two commands to create a "pool" and filesystem)
Casper
--
Expressed in this posting are my opinions. They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.
| |
| Chris F.A. Johnson 2006-07-22, 1:22 am |
| On 2006-07-22, base60 wrote:
> Casper H.S. Dik wrote:
>
> Well, they choose to not use xml for their system .conf
> files etc., so that's more than one can say for Sun 
That's a good reason to stay away from it!
> Disclaimer: yes, I prefer AIX (almost anything) to Solaris
--
Chris F.A. Johnson, author <http://cfaj.freeshell.org>
Shell Scripting Recipes: A Problem-Solution Approach (2005, Apress)
===== My code in this post, if any, assumes the POSIX locale
===== and is released under the GNU General Public Licence
| |
| base60 2006-07-22, 1:22 am |
| Chris F.A. Johnson wrote:
> On 2006-07-22, base60 wrote:
>
> That's a good reason to stay away from it!
The ironic part is that the Sun guys used to make fun of the oddities
of AIX
| |
| Chris F.A. Johnson 2006-07-22, 1:22 am |
| On 2006-07-22, base60 wrote:
> Chris F.A. Johnson wrote:
>
> The ironic part is that the Sun guys used to make fun of the oddities
> of AIX
Perhaps I should have been more specific:
That's a good reason to stay away from AIX!
--
Chris F.A. Johnson, author <http://cfaj.freeshell.org>
Shell Scripting Recipes: A Problem-Solution Approach (2005, Apress)
===== My code in this post, if any, assumes the POSIX locale
===== and is released under the GNU General Public Licence
| |
| Dave Hinz 2006-07-22, 1:18 pm |
| On Sat, 22 Jul 2006 02:01:01 -0400, Chris F.A. Johnson <cfajohnson@gmail.com> wrote:
> On 2006-07-22, base60 wrote:
[vbcol=seagreen]
> Perhaps I should have been more specific:
> That's a good reason to stay away from AIX!
And yet, the funny thing is, once you've been doing it long enough the
differences just don't matter anymore. It's all Unix, be it AIX, HP-UX,
Solaris, RedHat, Suse, Debian, Ubuntu...it just doesn't matter. I _do_
feel a bit dirty after working on a SCO box though, I must admit, and I
try to avoid those, but that's for other than technical reasons of
course.
| |
| Casper H.S. Dik 2006-07-22, 1:18 pm |
| base60 <nobody@whitehouse.com> writes:
>This was only available under opensolaris until quite recently.
>And given sun's flip/flops on this, I think I'd wait a bit
>before using it.
Which flip/flops?
ZFS was announced for S10 but it was made clear that it would
not be in the initial release. Then it was announced for update 2
and lo and behold it is in update 2.
Casper
--
Expressed in this posting are my opinions. They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.
| |
| Casper H.S. Dik 2006-07-22, 1:18 pm |
| "Chris F.A. Johnson" <cfajohnson@gmail.com> writes:
> That's a good reason to stay away from it!
The XML files are not for editing they merely describe how
the system is configured. Learning XML is only required if you
want to provide your own SMF enabled services.
Casper
| |
| base60 2006-07-22, 1:18 pm |
| Chris F.A. Johnson wrote:
> On 2006-07-22, base60 wrote:
>
> Perhaps I should have been more specific:
>
> That's a good reason to stay away from AIX!
LOL
Well, if you believe having your boot and config scripts in XML,
you're welcome to it.
| |
| base60 2006-07-22, 1:18 pm |
| Casper H.S. Dik wrote:
> base60 <nobody@whitehouse.com> writes:
>
>
>
> Which flip/flops?
>
> ZFS was announced for S10 but it was made clear that it would
> not be in the initial release. Then it was announced for update 2
> and lo and behold it is in update 2.
I'm simply repeating what I was told by Sun regarding ZFS... and
from what I've heard, others in this area were told the same.
Irregardless, it was just released and no-one with a real business
would fire it up for something they cared about until the dust
settled out.
| |
| base60 2006-07-22, 1:18 pm |
| Casper H.S. Dik wrote:
> "Chris F.A. Johnson" <cfajohnson@gmail.com> writes:
>
>
> The XML files are not for editing they merely describe how
> the system is configured.
And if you merely wish to change something or understand what
it is that Sun has done, you are now forced to do so in xml
rather than a shell.
Possibly it's just a personal bias, but I fail to see any
value-add on this.
The only real business Sun has is computers and they seem
to have figured out that Linux was making inroads at the
bottom and AIX at the top.
Sun's business is more or less stagnant and for that reason,
their stock is stalled at or under $5.00 -- much under at
the moment and dropping steadily.
Sol 10 was mostly about marketing and money.
Yes, ZFS is a VERY long overdue good thing... it's only
taken them, what, 16 years to after AIX and 8 - 10 after
HP-UX?
And then there was Openwin, which Sun would probably still
be based on if it wasn't for selling to the Feds.
If you're happy with Solaris and Sun, but all means, work
with them and stay with them...
However, judging from what is happening to their stock,
I'd say I have a fair amount of company in thinking that
Sol 10 was a mistake.
And, since this appears to be degrading the way these
threads usually do and I have no quarrel with you, I may
read your responses, but I won't reply.
Regards.
| |
| Chris F.A. Johnson 2006-07-22, 7:20 pm |
| On 2006-07-22, base60 wrote:
> Casper H.S. Dik wrote:
>
> And if you merely wish to change something or understand what
> it is that Sun has done, you are now forced to do so in xml
> rather than a shell.
>
> Possibly it's just a personal bias, but I fail to see any
> value-add on this.
I agree. I thought, from the earlier post, that AIX was the one
using XML.
--
Chris F.A. Johnson, author <http://cfaj.freeshell.org>
Shell Scripting Recipes: A Problem-Solution Approach (2005, Apress)
===== My code in this post, if any, assumes the POSIX locale
===== and is released under the GNU General Public Licence
| |
| Chris F.A. Johnson 2006-07-22, 7:20 pm |
| On 2006-07-22, Casper H.S Dik wrote:
> "Chris F.A. Johnson" <cfajohnson@gmail.com> writes:
>
>
> The XML files are not for editing they merely describe how
> the system is configured. Learning XML is only required if you
> want to provide your own SMF enabled services.
What is the purpose of making them more convoluted than necessary?
--
Chris F.A. Johnson, author <http://cfaj.freeshell.org>
Shell Scripting Recipes: A Problem-Solution Approach (2005, Apress)
===== My code in this post, if any, assumes the POSIX locale
===== and is released under the GNU General Public Licence
| |
| Mogens V. 2006-07-22, 7:20 pm |
| base60 wrote:
> Casper H.S. Dik wrote:
>
>
>
> And if you merely wish to change something or understand what
> it is that Sun has done, you are now forced to do so in xml
> rather than a shell.
> Dave Hintz:
>
> And yet, the funny thing is, once you've been doing it long enough
> the differences just don't matter anymore. It's all Unix, be it AIX,
> HP-UX, Solaris, RedHat, Suse, Debian, Ubuntu...it just doesn't matter.
So, the differences don't matter? When administering various versions of
*nix, differences do matter. Yes, we may all get along with the various
ways of doing things, but it doesn't make life easier if well worked
practices gets substituted for meaningless stupidities.
In the Linux world, those differences means extra work every time one
has has to get familiar with yet another distro line.
I'd much rather see basic *nix principles applied over the line, than
having yet another wrapper layer implemented.
GUI tooks, as an example, can be great, but fixing things through SSH on
a thingy where most everything gets autogenerated... XML then? Geez!
I haven't used Sol 10 as yet, but I'm sure I'd hate having to do admin
based on XML.
--
Kind regards,
Mogens V.
| |
| Dave Hinz 2006-07-23, 1:20 am |
| On Sat, 22 Jul 2006 22:06:13 +0200, Mogens V. <mogensv@NOSPAMvip.cybercity.dk> wrote:
> base60 wrote:
[vbcol=seagreen]
> So, the differences don't matter?
At a fundamental level, no, they don't.
> When administering various versions of
> *nix, differences do matter. Yes, we may all get along with the various
> ways of doing things, but it doesn't make life easier if well worked
> practices gets substituted for meaningless stupidities.
> In the Linux world, those differences means extra work every time one
> has has to get familiar with yet another distro line.
But if you understand the what, and don't focus on the how, it's still
all the same - just implemented differently at the end. The strategy
and tactics are the same.
> I'd much rather see basic *nix principles applied over the line, than
> having yet another wrapper layer implemented.
> GUI tooks, as an example, can be great, but fixing things through SSH on
> a thingy where most everything gets autogenerated... XML then? Geez!
> I haven't used Sol 10 as yet, but I'm sure I'd hate having to do admin
> based on XML.
I took the "new features in Solaris 10" class; the XML thing isn't
nearly as intrusive as you seem to think.
| |
| Tim Skirvin 2006-07-23, 1:20 am |
| Dave Hinz <DaveHinz@gmail.com> writes:
> I took the "new features in Solaris 10" class; the XML thing isn't
> nearly as intrusive as you seem to think.
Agreed. It's caused no problems for me whatsoever.
- Tim Skirvin (tskirvin@killfile.org)
--
http://www.killfile.org/~tskirvin/ Skirv's Homepage <FISH>< <*>
http://www.ks.uiuc.edu/~tskirvin/ Skirv's Work + CV
| |
| Logan Shaw 2006-07-23, 1:20 am |
| Chris F.A. Johnson wrote:
> On 2006-07-22, Casper H.S Dik wrote:
>
> What is the purpose of making them more convoluted than necessary?
That question contains the assumption that they are, in fact, more
convoluted than necessary.
I haven't used Solaris 10 extensively, but I assume the reasoning
went something like this: first, they decided to add features to
the startup process so that, rather than init scripts being a full
ordering based on numbers like "S10" and "S25", they would be a
partial ordering, thus allowing the system to do something like a
topological sort and intelligently start multiple things at once
if it is legitimate to do that. And they planned to add some
other features, like the ability to restart services that fail
(something that neither BSD-style rc.local nor System-V-style init
scripts support).
And they were faced with a question. These new features would
require a new language to describe the relationship of services
and so on, so should they:
(a) invent yet another custom syntax for the configuration files, or
(b) use XML?
And they chose XML, I've always assumed, because XML's whole purpose
is to factor out the custom syntaxes from the languages that every
different config file uses (and a zillion other types of languages).
So, personally, I don't think editing XML with vi is all that hard,
and I don't really think it's a big hardship to have some XML files
when the alternative is some proprietary syntax which isn't any
easier to deal with (given the advantage of having a really basic
knowledge of XML).
- Logan
| |
| Logan Shaw 2006-07-23, 1:20 am |
| base60 wrote:
> Casper H.S. Dik wrote:
>
> I'm simply repeating what I was told by Sun regarding ZFS... and
> from what I've heard, others in this area were told the same.
Actually, no, you're claiming that Sun has flip-flopped without
actually repeating anything they've said about it.
I do think some of the marketing campaign for the initial release
of Solaris 10 made it sound like it would have ZFS in it, but
then when you took a look at the documentation, you could tell
that wasn't really the case.
> Irregardless, it was just released and no-one with a real business
> would fire it up for something they cared about until the dust
> settled out.
I think you mean "regardless", but yeah, I agree with that. I have
that policy about every new filesystem ever, no matter what. In fact,
I generally believe that about any new technology at all. The
people who have really, really important data are going to run an
operating system released more than a year ago, and they're going
to be careful about what patches they apply to it as well. That's
not Solaris-specific; that's just good common sense. Well, maybe
not particular common, so maybe it's just good sense.
- Logan
| |
| Casper H.S. Dik 2006-07-23, 7:24 am |
| base60 <nobody@whitehouse.com> writes:
>And if you merely wish to change something or understand what
>it is that Sun has done, you are now forced to do so in xml
>rather than a shell.
No; you are not supposed to edit any of the XML files.
All configuration changes should be done through the
command line utilities provided (svccfg in particular)
>Sol 10 was mostly about marketing and money.
Really? More new features were added to Solaris 10
than any of Solaris 2.0 through 9 *combined*.
>Yes, ZFS is a VERY long overdue good thing... it's only
>taken them, what, 16 years to after AIX and 8 - 10 after
>HP-UX?
The competition has nothing like ZFS.
>And then there was Openwin, which Sun would probably still
>be based on if it wasn't for selling to the Feds.
And OpenWindows was a fine window system.
CDE and GNOME are just progressively worse.
Casper
--
Expressed in this posting are my opinions. They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.
| |
| Casper H.S. Dik 2006-07-23, 7:24 am |
| Logan Shaw <lshaw-usenet@austin.rr.com> writes:
>I haven't used Solaris 10 extensively, but I assume the reasoning
>went something like this: first, they decided to add features to
>the startup process so that, rather than init scripts being a full
>ordering based on numbers like "S10" and "S25", they would be a
>partial ordering, thus allowing the system to do something like a
>topological sort and intelligently start multiple things at once
>if it is legitimate to do that. And they planned to add some
>other features, like the ability to restart services that fail
>(something that neither BSD-style rc.local nor System-V-style init
>scripts support).
Plus the ability to use the expressed dependencies to
restart systems depending on a failed but restarted system or
to prevent services from starting if they depend on other
services which failed to start.
>And they were faced with a question. These new features would
>require a new language to describe the relationship of services
>and so on, so should they:
>(a) invent yet another custom syntax for the configuration files, or
>(b) use XML?
>And they chose XML, I've always assumed, because XML's whole purpose
>is to factor out the custom syntaxes from the languages that every
>different config file uses (and a zillion other types of languages).
And you need only one parser.
>So, personally, I don't think editing XML with vi is all that hard,
>and I don't really think it's a big hardship to have some XML files
>when the alternative is some proprietary syntax which isn't any
>easier to deal with (given the advantage of having a really basic
>knowledge of XML).
As I said before; the XML files provided with the OS are *not*
configuration files; they are supposed to be as immutable as
binaries and shell scripts delivered with the OS.
Not until you develop your own services you should need to edit
XML files or when you want to apply a site-specific profile.
Casper
--
Expressed in this posting are my opinions. They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.
| |
| Mogens V. 2006-07-23, 7:24 am |
| Tim Skirvin wrote:
> Dave Hinz <DaveHinz@gmail.com> writes:
>
>
>
>
> Agreed. It's caused no problems for me whatsoever.
Good to know. Why not? Good admin tools, or..?
--
Kind regards,
Mogens V.
| |
| Mogens V. 2006-07-23, 7:24 am |
| Casper H.S. Dik wrote:
> Not until you develop your own services you should need to edit
> XML files or when you want to apply a site-specific profile.
I'm thinking (it happens)...
On Linux/BSD, ever so often, I've had to implement my own startup
scripts, which has never been a problem on Redhat, Slackware or BSD's.
The reasons has been mostly because some server app wasn't available
with the type of init system used in the nix I was using, the particular
init script was inferior in design, or I was creating my own service,
like firewalling. needing a custom init script.
Now, Solaris may easily be better equipped with good init services, at
least that's what I've seen in my [admittedly] limited use of Sol8/9,
meaning I may not need to do this, at least only in rare cases.
Comments?
If one has to work out some XML init script, is it vastly more
complicated digging out the inner workings of existing ones to
understand what's going on?
I've done this on [non-XML] RH/Slack/Gentoo/BSD.. a nof times.
--
Kind regards,
Mogens V.
| |
| Dave Hinz 2006-07-23, 1:21 pm |
| On 23 Jul 2006 09:39:23 GMT, Casper H.S Dik <Casper.Dik@Sun.COM> wrote:
> base60 <nobody@whitehouse.com> writes:
>
>
> Really? More new features were added to Solaris 10
> than any of Solaris 2.0 through 9 *combined*.
See that's the sort of thing that annoys me (him, not you, Casper).
The guy hasn't done the most basic of research, apparently, and
dismisses it out of ignorance while pretending to understand it. Let's
save that kind of crap for the .advocacy groups OK base60?
Another statement made out of ignorance.
[vbcol=seagreen]
> The competition has nothing like ZFS.
Well, if you combine 2 or 3 things together, then sort of. But as an
entity, no. That said, I'm still not ready to use it on prod data, too
early. Dev? Sure. How else are you going to get experience and shake
out any bugs, or show that it's reliable in your environment?
| |
| Tim Skirvin 2006-07-23, 1:21 pm |
| "Mogens V." <mogensv@NOSPAMvip.cybercity.dk> writes:
> Good to know. Why not? Good admin tools, or..?
Mostly because I run a generalist-shop, so a new OS isn't very
scary.
There was a little bit of a learning-curve as I figured out what
services had been added that I wanted to turn off. And it was a little
bit intimidating at first when I heard that there was going to be a new
system... I thought that my old, interesting rc.d scripts were going to
stop working, which was not the case at all.
The new system is significantly slower *on the first boot*, as it
works out the relationships of all the SMF services; but on later boots,
it's significantly faster than the old S9 systems. Having heard an
explanation of what's happening and why from a Solaris engineer or two,
I'm not shocked that this is the case. And I wish that other companies
would follow Sun's lead on this; it's a *wonderful* advantage.
For the curious, this is the script I use to secure all of my
systems, S9 and S10, on initial install:
#!/bin/sh
echo "Disabling services"
for i in ssh sendmail
do
/usr/sbin/svcadm disable $i
done
for i in S88sendmail S47pppd S85power S72slpd S77dmi S80lp S80spc \
S95IIim S90wbem S82initsma S76snmpdx S77dmi S34dhcp
do
if [ -f /etc/rc2.d/$i ] ; then
/etc/rc2.d/$i stop
mv /etc/rc2.d/$i /etc/rc2.d/no$i
fi
if [ -f /etc/rc3.d/$i ] ; then
/etc/rc3.d/$i stop
mv /etc/rc3.d/$i /etc/rc2.d/no$i
fi
done
(I run a separate SSH daemon elsewhere.)
- Tim Skirvin (tskirvin@ks.uiuc.edu)
--
Theoretical and Computational http://www.ks.uiuc.edu/~tskirvin/
Biophysics, Beckman Institute, UIUC Senior Systems Administrator
| |
| Tim Skirvin 2006-07-23, 1:21 pm |
| "Mogens V." <mogensv@NOSPAMvip.cybercity.dk> writes:
> I'm thinking (it happens)...
> On Linux/BSD, ever so often, I've had to implement my own startup
> scripts, which has never been a problem on Redhat, Slackware or BSD's.
You can still use your old-style init.d scripts on Solaris. They
just don't get to take advantage the active-monitoring thing that the SMF
init stuff gets.
- Tim Skirvin (tskirvin@ks.uiuc.edu)
--
Theoretical and Computational http://www.ks.uiuc.edu/~tskirvin/
Biophysics, Beckman Institute, UIUC Senior Systems Administrator
| |
| base60 2006-07-23, 7:20 pm |
| Dave Hinz wrote:
> On 23 Jul 2006 09:39:23 GMT, Casper H.S Dik <Casper.Dik@Sun.COM> wrote:
OK..??
Possibly we're having a disconnect: I see nothing contradictory
between my statement and the above.
If you re-read, you should note that my complaint is the *number*
and type of changes
So, we would seem to agree regarding the number, anyway.
Sun, like IBM, HP, MS, etc., is in business to make money. Mgmt
dictates direction based upon what they think will generate the
most revenue for the company.
If you wish to believe something other than that, I have no intention
of attempting to dissuade you.
However, you can easily find how the market views recent Sun
innovations. If you believe the market is incorrect, you should go
long on Sun and then retire to the beach when it returns to it's
halcyon days of yore.
http://finance.yahoo.com
Disclaimer: I do not currently own any tech stocks, nor do I have any
direct or indirect financial interest in Sun or any of it's competitors.
[vbcol=seagreen]
>
> See that's the sort of thing that annoys me (him, not you, Casper).
> The guy hasn't done the most basic of research, apparently, and
> dismisses it out of ignorance while pretending to understand it. Let's
> save that kind of crap for the .advocacy groups OK base60?
Eh?
Pardon me for noting it, but you really didn't say anything above
except that I seem to piss you off.
Kindly note that the only thing that I'm "advocating" is that I
hope Sun has a fallback position.
If that is cause for you to get your panties in a wad, so be it.
>
>
> Another statement made out of ignorance.
Ah, well, that must be why Veritas FS has such a large installation
base on AIX, HPUX etc.
>
Possibly.
It may indeed be the best thing since sliced bread, and
after it's had a year or two to shake out and we find that
Sun still supports it, we may give it a shot for some of
our smaller production apps.
[vbcol=seagreen]
>
> Well, if you combine 2 or 3 things together, then sort of. But as an
> entity, no. That said, I'm still not ready to use it on prod data, too
> early.
Say it ain't so?
Surely, after your little rants, you don't want to find yourself
agreeing with me?
| |
| base60 2006-07-23, 7:20 pm |
| Logan Shaw wrote:
> base60 wrote:
>
> Actually, no, you're claiming that Sun has flip-flopped without
> actually repeating anything they've said about it.
Give some thought to reviewing a thread into which you've jumped
prior to commenting.
>
> I do think some of the marketing campaign for the initial release
> of Solaris 10 made it sound like it would have ZFS in it, but
> then when you took a look at the documentation, you could tell
> that wasn't really the case.
Given the choice of a flip-flop or misleading marketing, I think
the kinder interpretation would be the former.
>
>
> I think you mean "regardless",
No, I meant irregardless. Its use is colloquial and common
in informal writing... and this is informal.
And grammar comments invariably come back to haunt;
> but yeah, I agree with that. I have
but, yeah, I agree with that. :-)
> that policy about every new filesystem ever, no matter what. In fact,
> I generally believe that about any new technology at all. The
> people who have really, really important data are going to run an
> operating system released more than a year ago, and they're going
> to be careful about what patches they apply to it as well. That's
> not Solaris-specific; that's just good common sense. Well, maybe
> not particular common, so maybe it's just good sense.
>
> - Logan
| |
| Casper H.S. Dik 2006-07-24, 1:21 am |
| "Mogens V." <mogensv@NOSPAMvip.cybercity.dk> writes:
>Casper H.S. Dik wrote:
[vbcol=seagreen]
>I'm thinking (it happens)...
>On Linux/BSD, ever so often, I've had to implement my own startup
>scripts, which has never been a problem on Redhat, Slackware or BSD's.
If you must, the old rc.d scripting method still works.
>If one has to work out some XML init script, is it vastly more
>complicated digging out the inner workings of existing ones to
>understand what's going on?
The XML is not a "script"; the start/stop scripts are generally
unchanged; the XML file contains the service's properties, their
default values, dependencies and such.
Casper
--
Expressed in this posting are my opinions. They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.
| |
| Dave Hinz 2006-07-24, 1:24 pm |
| On Sun, 23 Jul 2006 19:12:38 GMT, base60 <nobody@whitehouse.com> wrote:
> Dave Hinz wrote:
[vbcol=seagreen]
> If you re-read, you should note that my complaint is the *number*
> and type of changes
How can you meaningfully count the number of _anything_ that changed in
software? You can count it any way you want, it doesn't mean a thing.
> So, we would seem to agree regarding the number, anyway.
Do we? I lost track.
> Sun, like IBM, HP, MS, etc., is in business to make money. Mgmt
> dictates direction based upon what they think will generate the
> most revenue for the company.
> If you wish to believe something other than that, I have no intention
> of attempting to dissuade you.
The corporation, sure, is in business to make money. Believe it or not,
the engineers and so on who design the OS are doing it because, in part,
they want to do it well. If that isn't obvious to you, well, nothing
I'm going to say will change that.
> However, you can easily find how the market views recent Sun
> innovations. If you believe the market is incorrect, you should go
> long on Sun and then retire to the beach when it returns to it's
> halcyon days of yore. > http://finance.yahoo.com
Yes, I'm familiar with the stock price history. It's also astonishingly
irrelevant to a discussion of if some guy should use AIX or Solaris for
his new project.
> Disclaimer: I do not currently own any tech stocks, nor do I have any
> direct or indirect financial interest in Sun or any of it's competitors.
I do, on both counts.
[vbcol=seagreen]
> Eh?
> Pardon me for noting it, but you really didn't say anything above
> except that I seem to piss you off.
Piss me off? Hardly. It's just that "my dog is better than your dog"
just doesn't accomplish anything. I'm sure there's a place where it's
more on topic and you could find an enthusiastic group of participants,
though.
> Kindly note that the only thing that I'm "advocating" is that I
> hope Sun has a fallback position.
> If that is cause for you to get your panties in a wad, so be it.
There's apparently someone here with undie bundleage, yes, but it's not
on this side of the conversation.
[vbcol=seagreen]
> Ah, well, that must be why Veritas FS has such a large installation
> base on AIX, HPUX etc.
I'll take "15 years of history" for 500 please, Trebek.
>
> Possibly.
> It may indeed be the best thing since sliced bread, and
> after it's had a year or two to shake out and we find that
> Sun still supports it, we may give it a shot for some of
> our smaller production apps.
Great. Hope it works out for you as well as it looks like it may.
[vbcol=seagreen]
> Say it ain't so?
I'm not risking my employer's data on _anything_ brand new. Nothing
unique to Sun or ZFS. Deciding on what to do based on how shiny it is,
is poor design.
> Surely, after your little rants, you don't want to find yourself
> agreeing with me?
Um, rants? Are you sure we're reading the same group?
| |
| Stuart Krivis 2006-07-24, 1:24 pm |
| On 22 Jul 2006 13:21:48 GMT, Dave Hinz <DaveHinz@gmail.com> wrote:
>Solaris, RedHat, Suse, Debian, Ubuntu...it just doesn't matter. I _do_
>feel a bit dirty after working on a SCO box though, I must admit, and I
>try to avoid those, but that's for other than technical reasons of
>course.
I have to work with OpenServer boxes. It feels 10 years behind
everyone else. Then again, the software we use on them is running
under a RSTS emulator. :-)
We have a lot of Suns, and I've worked with Solaris since '96 or so.
Sun has always seemed to offer a pretty good solution overall.
Sometimes Sun is superior, sometimes not, but it's always at least in
the running.
We just got some big Power 5 boxes, and the hardware looks really
nice. I guess I'll be seeing what AIX is like. I'm betting it will be
better in some ways and in the running on others. :-)
| |
| Doug Freyburger 2006-07-24, 7:22 pm |
| Dave Hinz wrote:
> Mogens V. wrote:
>
>
>
> At a fundamental level, no, they don't.
I learned to drive a car with an automatic transmission driving on
the port side of the road. Does that mean I'll never be able to
drive a manual transmission? I got in one, puzzled out the controls
a while and started driving. In under a day my brain had a working
program for it. I went to Japan where they drive on the starboard
side of the road and I couldn't handle it, right? Nah, I got in a van,
puzzled out the controls a while and started driving very cautiously.
It took a week before I could manuever fluently in traffic but there
I was carrying parts from point A to point B.
I learned to drive a bicycle when I was a kid. Wanna bet I could
get on a motorcycle, leave through the manual, puzzle out the
controls a while, and apply my car driving and bicycle driving
skills together, and get to point B?
At a fundamental level it's all driving. At a fundamental level it's
all
administering UNIX hosts and for that matter managing data flow
through data centers.
Specialization is the path to extinction in both evolution of species
and careers. If you don't think you can puzzle through the controls
for a bit and be working on the next version of UNIX then you have
specialized enough to endanger your career.
| |
|
| Begin <1153776226.221257.269030@s13g2000cwa.googlegroups.com>
On 2006-07-24, Doug Freyburger <dfreybur@yahoo.com> wrote:
> I learned to drive a car with an automatic transmission driving on
> the port side of the road. Does that mean I'll never be able to
> drive a manual transmission? I got in one, puzzled out the controls
> a while and started driving. In under a day my brain had a working
> program for it.
Maybe you can, but looking at the standards of driving tests and the
percentages of faillures, lots of people do have trouble figuring out
new things, or even how to do things they supposedly know how to and
have done for a considerable time, well, safely, etcetera.
There's a reason many jurisdictions restrict holders of licences
obtained driving an automatic from driving a manual unless and until
they pass the test in a manual. Unnecessairy for you, perhaps, but for
all the driving equivalent of windows users Out There, apparently not.
> I went to Japan where they drive on the starboard side of the road
> and I couldn't handle it, right? Nah, I got in a van, puzzled out the
> controls a while and started driving very cautiously. It took a week
> before I could manuever fluently in traffic but there I was carrying
> parts from point A to point B.
So far, so good. Recently there was a nasty accident in the UK, caused
by a German cop on vacation who, taking a turn, swerved on the wrong
lane out of habit. He thought he'd figured it out, habit took over,
accident. Head-on crash with a motorcyclist.[0]
> Specialization is the path to extinction in both evolution of species
> and careers. If you don't think you can puzzle through the controls
> for a bit and be working on the next version of UNIX then you have
> specialized enough to endanger your career.
Some people appear to be learn-once-never-again types. Don't ask me why,
I don't understand it either. Except perhaps that laziness and habit
would be awfully atractive if it wasn't so boring.
[0] http://news.bbc.co.uk/2/hi/uk_news/...ral/5198752.stm
--
j p d (at) d s b (dot) t u d e l f t (dot) n l .
This message was originally posted on Usenet in plain text.
Any other representation, additions, or changes do not have my
consent and may be a violation of international copyright law.
| |
| Doug Freyburger 2006-07-28, 1:24 pm |
| jpd wrote:
> Doug Freyburger wrote:
>
>
> Maybe you can, but looking at the standards of driving tests and the
> percentages of faillures, lots of people do have trouble figuring out
> new things, or even how to do things they supposedly know how to and
> have done for a considerable time, well, safely, etcetera.
> ...
>
> Some people appear to be learn-once-never-again types. Don't ask me why,
> I don't understand it either. Except perhaps that laziness and habit
> would be awfully atractive if it wasn't so boring.
Lesson to be learned - SysAdmining is *not* for everyone. To ever
get good in the field it takes a learn-many-times approach.
In the simple case of someone having trouble switching from an
automatic to a manual transmission, anyone having troubles with
machines on that level likely shouldn't be in any sector of IT at
all. "I saw a news article that pay is good in IT so I applied" is
NOT a reason to be in the field. Yes, I am quite elitist. ;^)
|
|
|
|
|