|
Home > Archive > Unix administration > January 2004 > How often should I reboot Solaris and LynxOS
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]
| Author |
How often should I reboot Solaris and LynxOS
|
|
|
| Hi,
Would someone please give me some pointers on a trend analysis on resource
leaks for Solaris 9 and LynxOS (or in general on UNIX machines). Basically,
I need statistics to determine how often I need to restart these machines to
avoid unplanned failures and keep Heisenbugs at bay. Thanks.
ekl
| |
| davehinz@spamcop.net 2004-01-23, 4:29 pm |
| In comp.unix.admin EKL <En-Kuang_Lung@raytheon.com> wrote:quote:
> Hi,
quote:
> Would someone please give me some pointers on a trend analysis on resource
> leaks for Solaris 9 and LynxOS (or in general on UNIX machines).
(I assume you mean "Linux"; if "LynxOS" is something that I haven't
heard of, then I can't comment on that).
Reboots "just because" aren't needed - most of my boxes end up getting
rebooted somewhere around a year or sooner, though. Usually, it's
a case of needing to do a kernel patch, or a physical change to the
hardware.
quote:
> Basically,
> I need statistics to determine how often I need to restart these machines to
> avoid unplanned failures and keep Heisenbugs at bay. Thanks.
While it does happen from time to time, it's an exception rather
than the rule. Uptimes approaching and exceeding a year aren't all
that remarkable. A daily, weekly, or monthly reboot "just because"
is a political response to a non-technical issue, if anything.
Hope this helps,
Dave Hinz
| |
| Hubert Feyrer 2004-01-23, 4:29 pm |
| In comp.unix.internals EKL <En-Kuang_Lung@raytheon.com> wrote:quote:
> Would someone please give me some pointers on a trend analysis on resource
> leaks for Solaris 9 and LynxOS (or in general on UNIX machines). Basically,
> I need statistics to determine how often I need to restart these machines to
> avoid unplanned failures and keep Heisenbugs at bay. Thanks.
We had Solaris 7 running on a Dell PC (Optiplex GX1p or so) as a student
login box. It was used for simple things like email, news-reading and irc,
and screen sessions. Nothing too hard. Besides that, it was also used for
a "system programming" course with socket programming, shared memory,
semaphores, fork bombs and other stuff that stresses the internals a
fair bit.
The machine at an uptime of >2 years(!) when someone popped a Zip diskette
into it and we had to find out the zip driver likes to panic. X-)
(Pic at http://www.feyrer.de/Texts/Troja/tabaluga-2y.jpg).
I guess so much for stability of Solaris. And Sol9/sparc sure isn't worse
than Sol7/x86 (IF you don't have flaky... that's a different story that
I'll not go into now, just one word of warning: avoide 750MHz CPUs in
SunFire 280Rs, they're known to fail often!)
- Hubert
--
Want to get a clue on IPv6 but don't know where to start? Try this:
* Basics -> http://www.onlamp.com/pub/a/onlamp/...6_tutorial.html
* Setup -> http://www.onlamp.com/pub/a/onlamp/...6_tutorial.html
Of course with your #1 IPv6 ready operating system -> http://www.NetBSD.org/
| |
| Kent Wormsdorf 2004-01-23, 4:29 pm |
| d. 24/07/03 16:35 skrev EKL på En-Kuang_Lung@raytheon.com i artiklen
x8STa.3815$c6.3313@bos-service2.ext.raytheon.com:
quote:
> Hi,
>
> Would someone please give me some pointers on a trend analysis on resource
> leaks for Solaris 9 and LynxOS (or in general on UNIX machines). Basically,
> I need statistics to determine how often I need to restart these machines to
> avoid unplanned failures and keep Heisenbugs at bay. Thanks.
>
> ekl
>
>
>
Hi,
I only shut my Sun Blade 100 down for the night, because it is placed right
beside my bed. Else I wouldn't shut it down at all. (But I can't sleep
because of the noice it generates.)
The only resource Hog I have is netscape 6.x, wich I restart a couple of
times each day.
Greetings from Denmark
Kent Wormsdorf
--
PS. (Sorry, only in danish.)
Er der nogen der mangler en mand med lidt viden om Sun HW og SW,
så må I gerne sende mig en E-mail.
Jeg er uddannet datamatiker og har senest arbejdet som teknisk presale
konsulent.
| |
| Ben Elliston 2004-01-23, 4:29 pm |
| William Ahern <william@wilbur.25thandClement.com> writes:
quote:
> *nix systems are just that, a _system_. They are ecosystems of
> software. If one part of the system is sick, you don't shutdown the
> rest of the ecosystem. For what, spite? Ignorance? *nix systems are
> about process and procedure, aided by technology (not technology
> aided by process and procedure, as in rebooting every few days).
This painted a funny image in my mind of a doctor throwing a patient
reporting a sore throat to the floor and applying the "cardiac arrest
paddles" :-)
Ben
| |
|
| EKL wrote:quote:
> Hi,
>
> Would someone please give me some pointers on a trend analysis on resource
> leaks for Solaris 9 and LynxOS (or in general on UNIX machines). Basically,
> I need statistics to determine how often I need to restart these machines to
> avoid unplanned failures and keep Heisenbugs at bay. Thanks.
>
> ekl
>
>
>
I would recommend, if you can, rebooting after any significant
configuration changes have been made. That way, you get to confirm that
the system reboots correctly so if some hardware problem or power
outage occurs, then you have a good chance that the system will come
back up without intervention.
Others have mentioned having to reboot for some patches etc.
Pete
| |
| Hubert Feyrer 2004-01-23, 4:29 pm |
| In comp.unix.internals EKL <En-Kuang_Lung@raytheon.com> wrote:quote:
> Would someone please give me some pointers on a trend analysis on resource
> leaks for Solaris 9 and LynxOS (or in general on UNIX machines). Basically,
> I need statistics to determine how often I need to restart these machines to
> avoid unplanned failures and keep Heisenbugs at bay. Thanks.
We had Solaris 7 running on a Dell PC (Optiplex GX1p or so) as a student
login box. It was used for simple things like email, news-reading and irc,
and screen sessions. Nothing too hard. Besides that, it was also used for
a "system programming" course with socket programming, shared memory,
semaphores, fork bombs and other stuff that stresses the internals a
fair bit.
The machine at an uptime of >2 years(!) when someone popped a Zip diskette
into it and we had to find out the zip driver likes to panic. X-)
(Pic at http://www.feyrer.de/Texts/Troja/tabaluga-2y.jpg).
I guess so much for stability of Solaris. And Sol9/sparc sure isn't worse
than Sol7/x86 (IF you don't have flaky... that's a different story that
I'll not go into now, just one word of warning: avoide 750MHz CPUs in
SunFire 280Rs, they're known to fail often!)
- Hubert
--
Want to get a clue on IPv6 but don't know where to start? Try this:
* Basics -> http://www.onlamp.com/pub/a/onlamp/...6_tutorial.html
* Setup -> http://www.onlamp.com/pub/a/onlamp/...6_tutorial.html
Of course with your #1 IPv6 ready operating system -> http://www.NetBSD.org/
| |
| Ben Elliston 2004-01-23, 4:29 pm |
| William Ahern <william@wilbur.25thandClement.com> writes:
quote:
> *nix systems are just that, a _system_. They are ecosystems of
> software. If one part of the system is sick, you don't shutdown the
> rest of the ecosystem. For what, spite? Ignorance? *nix systems are
> about process and procedure, aided by technology (not technology
> aided by process and procedure, as in rebooting every few days).
This painted a funny image in my mind of a doctor throwing a patient
reporting a sore throat to the floor and applying the "cardiac arrest
paddles" :-)
Ben
| |
| Bill Vermillion 2004-01-23, 4:33 pm |
| In article <x8STa.3815$c6.3313@bos-service2.ext.raytheon.com>,
EKL <En-Kuang_Lung@raytheon.com> wrote:
quote:
>Would someone please give me some pointers on a trend analysis
>on resource leaks for Solaris 9 and LynxOS (or in general on
>UNIX machines). Basically, I need statistics to determine how
>often I need to restart these machines to avoid unplanned
>failures and keep Heisenbugs at bay. Thanks.
Some Unix OSes had a problem with a timer [which NT copied]that
would go negative and needed to be rebooted every 248 days.
But that was fixed in almost all I know of about 3 or 4 years ago.
I now wind up rebooting every 18 months to 2 years as I also upgade
the OS at that time.
--
Bill Vermillion - bv @ wjv . com
| |
| Bill Vermillion 2004-01-23, 4:33 pm |
| In article <x8STa.3815$c6.3313@bos-service2.ext.raytheon.com>,
EKL <En-Kuang_Lung@raytheon.com> wrote:
quote:
>Would someone please give me some pointers on a trend analysis
>on resource leaks for Solaris 9 and LynxOS (or in general on
>UNIX machines). Basically, I need statistics to determine how
>often I need to restart these machines to avoid unplanned
>failures and keep Heisenbugs at bay. Thanks.
Some Unix OSes had a problem with a timer [which NT copied]that
would go negative and needed to be rebooted every 248 days.
But that was fixed in almost all I know of about 3 or 4 years ago.
I now wind up rebooting every 18 months to 2 years as I also upgade
the OS at that time.
--
Bill Vermillion - bv @ wjv . com
| |
|
| > On Solaris, it's something you are advised to do after installingquote:
> certain types of patches.
Unfortunately on Sol8, the 'recommended patch cluster' requires
installing in single-user mode, and rebooting. So depending on how
up-to-date you want to be, will govern how often you have to reboot.
I haven't patched one of our sol9 machines yet, so I can't comment on
this. The sol8 patch cluster just recently started requesting su
mode, so I am guessing sol9 won't need it until it gets old too...
Other than that the combination of solaris and SPARC is like a rock.
More stable than the linux PC's we had (we really pound the machines
with I/O and CPU).
We recently replace 4cpu's in our 12-banger server (air-conditioning
went in server-room), and did not have to take the machine down!
Pretty neat.
~S
| |
|
| > On Solaris, it's something you are advised to do after installingquote:
> certain types of patches.
Unfortunately on Sol8, the 'recommended patch cluster' requires
installing in single-user mode, and rebooting. So depending on how
up-to-date you want to be, will govern how often you have to reboot.
I haven't patched one of our sol9 machines yet, so I can't comment on
this. The sol8 patch cluster just recently started requesting su
mode, so I am guessing sol9 won't need it until it gets old too...
Other than that the combination of solaris and SPARC is like a rock.
More stable than the linux PC's we had (we really pound the machines
with I/O and CPU).
We recently replace 4cpu's in our 12-banger server (air-conditioning
went in server-room), and did not have to take the machine down!
Pretty neat.
~S
| |
| Dennis Clarke 2004-01-23, 4:33 pm |
| Bill Vermillion wrote:[QUOTE][color=darkred]
> In article <x8STa.3815$c6.3313@bos-service2.ext.raytheon.com>,
> EKL <En-Kuang_Lung@raytheon.com> wrote:
>
>
I find a good power failure seems to work for me. 
Other than that, when you NEED to. Patches. Building is on fire.
That sort of thing.
Dennis
| |
| Dennis Clarke 2004-01-23, 4:33 pm |
| Bill Vermillion wrote:[QUOTE][color=darkred]
> In article <x8STa.3815$c6.3313@bos-service2.ext.raytheon.com>,
> EKL <En-Kuang_Lung@raytheon.com> wrote:
>
>
I find a good power failure seems to work for me.
Other than that, when you NEED to. Patches. Building is on fire.
That sort of thing.
Dennis
| |
| David Schwartz 2004-01-23, 4:33 pm |
|
"Dennis Clarke" <dclarke@blastwave.org> wrote in message
news:3F3D337D.8080108@blastwave.org...quote:
> Bill Vermillion wrote:
>
> I find a good power failure seems to work for me.
>
> Other than that, when you NEED to. Patches. Building is on fire.
> That sort of thing.
>
> Dennis
For more on this, check out this page:
http://www.daemon-systems.org/man/halt.8.html
Specifically the documentation about the '-n' option.
DS
| |
| David Schwartz 2004-01-23, 4:33 pm |
|
"Dennis Clarke" <dclarke@blastwave.org> wrote in message
news:3F3D337D.8080108@blastwave.org...quote:
> Bill Vermillion wrote:
>
> I find a good power failure seems to work for me.
>
> Other than that, when you NEED to. Patches. Building is on fire.
> That sort of thing.
>
> Dennis
For more on this, check out this page:
http://www.daemon-systems.org/man/halt.8.html
Specifically the documentation about the '-n' option.
DS
| |
|
|
|
|
| Cypherpunk@nyc.rr.com 2004-01-23, 4:33 pm |
| In comp.unix.solaris Philip Brown <phil+s3@bolthole.no-bots.com> wrote:quote:
> dclarke@blastwave.org wrote:
>
> Have you found it helps to reboot, when the building is on fire, Dennis?
My SPARC passed a year of being up, then
there was this little ol' power failure.
| |
| Cypherpunk@nyc.rr.com 2004-01-23, 4:33 pm |
| In comp.unix.solaris Philip Brown <phil+s3@bolthole.no-bots.com> wrote:quote:
> dclarke@blastwave.org wrote:
>
> Have you found it helps to reboot, when the building is on fire, Dennis?
My SPARC passed a year of being up, then
there was this little ol' power failure.
| |
| Anthony Mandic 2004-01-23, 4:33 pm |
| Cypherpoke@nyc.rr.com wrote:quote:
>
> In comp.unix.solaris Philip Brown <phil+s3@bolthole.no-bots.com> wrote:
>
> My SPARC passed a year of being up, then
> there was this little ol' power failure.
It didn't help your Macintrash, of course, Plugger, since "youz"
have to reboot it everyday anyway. Next time train your rabbits
to keep away from the power cords.
-am © 2003
| |
| Anthony Mandic 2004-01-23, 4:33 pm |
| Cypherpoke@nyc.rr.com wrote:quote:
>
> In comp.unix.solaris Philip Brown <phil+s3@bolthole.no-bots.com> wrote:
>
> My SPARC passed a year of being up, then
> there was this little ol' power failure.
It didn't help your Macintrash, of course, Plugger, since "youz"
have to reboot it everyday anyway. Next time train your rabbits
to keep away from the power cords.
-am © 2003
| |
| Anthony Mandic 2004-01-23, 4:33 pm |
| Cypherpunk@nyc.rr.com wrote:quote:
> Philip Brown <phil+s3@bolthole.no-bots.com> wrote:
>
> My SPARC passed a year of being up, then
> there was this little ol' power failure.
Let's talk about Star Trek!
Anthony Mandic <mailto:am@peppler.org>
<mailto:anthony@isug.com>
<mailto:amandic@start.com.au> wrote in
3DC1F128.A354F3FE@start.com.au:quote:
>
> As we know, the Borg didn't know about humans until Q introduced
> them to Enterprise (and vice versa).
>
> So, like 99.99% of science fiction, not much thought was given to
> logic and sense.
Why does anyone even write Sci-Fi? Huh? Inquiring minds wanna know!
Why does literature so often fail to include references to buttplugs?
I'm fascinated by buttplugs, and mention them all the time!
----
As it is, plugger's committed a CRIMINAL OFFENCE as far as the
law in Australia is concerned. It's ILLEGAL TO MAKE FUN OF ME,
dammit. I had a special law passed to PROTECT ME FROM MEANIES.
It won't take too much effort to get him EXTRADITED and become
someone's XXXXX at Her Majesty's pleasure.
I have an ACE up my SLEEVE!!!
I'll go to the POLICE and BURST into TEARS!!!
That'll teach everybody to mess with me.
-am © 2003
| |
| Anthony Mandic 2004-01-23, 4:33 pm |
| Cypherpunk@nyc.rr.com wrote:quote:
> Philip Brown <phil+s3@bolthole.no-bots.com> wrote:
>
> My SPARC passed a year of being up, then
> there was this little ol' power failure.
Let's talk about Star Trek!
Anthony Mandic <mailto:am@peppler.org>
<mailto:anthony@isug.com>
<mailto:amandic@start.com.au> wrote in
3DC1F128.A354F3FE@start.com.au:quote:
>
> As we know, the Borg didn't know about humans until Q introduced
> them to Enterprise (and vice versa).
>
> So, like 99.99% of science fiction, not much thought was given to
> logic and sense.
Why does anyone even write Sci-Fi? Huh? Inquiring minds wanna know!
Why does literature so often fail to include references to buttplugs?
I'm fascinated by buttplugs, and mention them all the time!
----
As it is, plugger's committed a CRIMINAL OFFENCE as far as the
law in Australia is concerned. It's ILLEGAL TO MAKE FUN OF ME,
dammit. I had a special law passed to PROTECT ME FROM MEANIES.
It won't take too much effort to get him EXTRADITED and become
someone's XXXXX at Her Majesty's pleasure.
I have an ACE up my SLEEVE!!!
I'll go to the POLICE and BURST into TEARS!!!
That'll teach everybody to mess with me.
-am © 2003
|
|
|
|
|