asterisk users - Sep 2005 - Asterisk 1.0.9 long term stability <--thread hijack, why not reboot?

Disclaimer: Not a troll

I'm curious as to this obsession with uptime is. All of the posts of this
type are along the lines of "After X days, Y thing does not work but if I
reload or reboot, it's OK" - so why not cron a reboot? Is it considered
bad
form or something like that? I reboot every night whether it is needed or
not, not afraid to admit it, and everything works fine for me. 

We also do the "Sunday reboot" of all of our Windows servers as well
as
restarting all of the critical services such as IIS , SQL, Exchange etc
nightly. It helps, a lot (Exchange is a notorious memory leaker)

Of course, if your install processes calls 24/7 that's a different story.
However, I expect that the majority of Asterisk installs are for a 9-to-5
type of operation. We run two shifts here, and we stop processing calls at
10 PM, and start again at about 6 AM - a large window of opportunity to
reboot. Why not take advantage of it?

I've also heard it said, something along the lines of: "If you have to
reboot, your server isn't set up correctly" to which I say piffle. Even
NASA
has rebooted the Mars probes after they land and I understand that they run
VXWorks, incidentally, the same RTOS that my Mitel 3300 uses, and *even
Mitel* recommends periodic reboots, which we duly cron every night, 2 AM. 

24/7/365 installs aside, is there a reason why reboots seem to be frowned
upon? Again, not trolling, just curious.

Yes.. because if a reboot is needed it isn't setup correctly.  
Reboots are a windows things.   A correctly setup Linux server should
never need rebooted.

On 9/14/05, Colin Anderson <ColinA@landmarkmasterbuilder.com>
wrote:
> Disclaimer: Not a troll
> 
> I'm curious as to this obsession with uptime is. All of the posts of
this
> type are along the lines of "After X days, Y thing does not work but
if I
> reload or reboot, it's OK" - so why not cron a reboot? Is it
considered bad
> form or something like that? I reboot every night whether it is needed or
> not, not afraid to admit it, and everything works fine for me.
> 
> We also do the "Sunday reboot" of all of our Windows servers as
well as
> restarting all of the critical services such as IIS , SQL, Exchange etc
> nightly. It helps, a lot (Exchange is a notorious memory leaker)
> 
> Of course, if your install processes calls 24/7 that's a different
story.
> However, I expect that the majority of Asterisk installs are for a 9-to-5
> type of operation. We run two shifts here, and we stop processing calls at
> 10 PM, and start again at about 6 AM - a large window of opportunity to
> reboot. Why not take advantage of it?
> 
> I've also heard it said, something along the lines of: "If you
have to
> reboot, your server isn't set up correctly" to which I say piffle.
Even NASA
> has rebooted the Mars probes after they land and I understand that they run
> VXWorks, incidentally, the same RTOS that my Mitel 3300 uses, and *even
> Mitel* recommends periodic reboots, which we duly cron every night, 2 AM.
> 
> 24/7/365 installs aside, is there a reason why reboots seem to be frowned
> upon? Again, not trolling, just curious.
> _______________________________________________
> --Bandwidth and Colocation sponsored by Easynews.com --
> 
> Asterisk-Users mailing list
> Asterisk-Users@lists.digium.com
> http://lists.digium.com/mailman/listinfo/asterisk-users
> To UNSUBSCRIBE or update options visit:
>    http://lists.digium.com/mailman/listinfo/asterisk-users
>

Just my opinions but....

A PBX should not require reboots EVER.  Traditional proprietary PBX's
don't
require reboots and if Asterisk does then it has problems and is not ready
to be a viable alternative IMHO.

I am sure that some people will say they know of some proprietary PBX's that
do require reboots.  My response would be that my 'expectation' for ANY
PBX
and even servers in general is NO REBOOTS.  I don't think that is
unreasonable.

This opinion has NOTHING to do with the whole Linux vs Windows Jihad.  That
is a whole other argument.  I am strictly talking about PBX's regardless of
if they are Linux or Windows or traditional proprietary or whatever.
Rebooting at 3am everyday just because you can should not justify permitting
the existance of problems that require it in the first place.  I hope that
makes sense. :)

Again, just my 2 cents.

-----Original Message-----
From: Colin Anderson [mailto:ColinA@landmarkmasterbuilder.com] 
Sent: Wednesday, September 14, 2005 9:34 AM
To: 'Asterisk Users Mailing List - Non-Commercial Discussion'
Subject: RE: [Asterisk-Users] Asterisk 1.0.9 long term stability
<--threadhijack, why not reboot?

Disclaimer: Not a troll

I'm curious as to this obsession with uptime is. All of the posts of this
type are along the lines of "After X days, Y thing does not work but if I
reload or reboot, it's OK" - so why not cron a reboot? Is it considered
bad
form or something like that? I reboot every night whether it is needed or
not, not afraid to admit it, and everything works fine for me. 

We also do the "Sunday reboot" of all of our Windows servers as well
as
restarting all of the critical services such as IIS , SQL, Exchange etc
nightly. It helps, a lot (Exchange is a notorious memory leaker)

Of course, if your install processes calls 24/7 that's a different story.
However, I expect that the majority of Asterisk installs are for a 9-to-5
type of operation. We run two shifts here, and we stop processing calls at
10 PM, and start again at about 6 AM - a large window of opportunity to
reboot. Why not take advantage of it?

I've also heard it said, something along the lines of: "If you have to
reboot, your server isn't set up correctly" to which I say piffle. Even
NASA
has rebooted the Mars probes after they land and I understand that they run
VXWorks, incidentally, the same RTOS that my Mitel 3300 uses, and *even
Mitel* recommends periodic reboots, which we duly cron every night, 2 AM. 

24/7/365 installs aside, is there a reason why reboots seem to be frowned
upon? Again, not trolling, just curious.

On Wednesday 14 September 2005 12:34, Colin Anderson
wrote:
> I'm curious as to this obsession with uptime is. All of the posts of
this
> type are along the lines of "After X days, Y thing does not work but
if I
> reload or reboot, it's OK" - so why not cron a reboot? Is it
considered bad
> form or something like that? I reboot every night whether it is needed or
> not, not afraid to admit it, and everything works fine for me.
Rebooting indicates that there is a problem that needs to be addressed.  It 
has nothing to do with uptime wars but with reliability and stability.  I 
don't care if the system's down for 3 minutes due to reboot, I am
concerned
that there is an underlying issue that you're merely masking by rebooting.
> We also do the "Sunday reboot" of all of our Windows servers as
well as
> restarting all of the critical services such as IIS , SQL, Exchange etc
> nightly. It helps, a lot (Exchange is a notorious memory leaker)
Exactly.  With closed-source proprietary applications you can't fix the 
problem so this is all you can do.  With Asterisk and OSS you can fix it and 
increase code quality.

Would you accept the need for a weekly reboot of your Nortel Option 11?  How 
about if your car required you to remove the battery for two minutes once a 
month?  Your VCR?  How about a clock radio that had to be unplugged once a 
week to fix weird little issues?

I put Asterisk into the same class as these kinds of devices.  They must be up 
and stay up.  If they need to be rebooted then there's an issue that needs
to
be addressed and corrected.
> I've also heard it said, something along the lines of: "If you
have to
> reboot, your server isn't set up correctly" to which I say piffle.
Even
> NASA has rebooted the Mars probes after they land and I understand that
> they run VXWorks, incidentally, the same RTOS that my Mitel 3300 uses, and
> *even Mitel* recommends periodic reboots, which we duly cron every night, 2
> AM.
Hmm, I guess I won't be buying any Mitel equipment.  The MARS rovers were 
designed to be totally shut down as a last measure to ensure everything is 
starting up as they'd simulated on Earth and that there was no high-energy 
radiation glitches due to space travel.

-A.

Great comments everyone thanks and thanks for not flaming me. 
>Rebooting indicates that there is a problem that needs to be addressed.  It
>has nothing to do with uptime wars but with reliability and stability.  I 
>don't care if the system's down for 3 minutes due to reboot, I am
concerned
>that there is an underlying issue that you're merely masking by
rebooting.
In my case, I have the TDM04 static problem and that is the sole reason why
I reboot.
Granted, I'm running 1.0-stable and I understand it's fixed in -head or
1.2
or whatever 
but I don't want to go through the grief of upgrading to fix it and find out
that other issues are introduced. A lot of guys say "upgrade, upgrade
it's
way better" but then I read posts like the gentleman this morning who has
his console going crazy when he upgraded. My plan is to wait out the rest 
of the year until 1.2 has a few months under it's belt then upgrade (hey,
isn't today the release day?). 
Until then, reboot to me is a perfectly acceptable alternative and it does
not impact business operations one whit. You may feel differently, and yes,
I concede that I am masking an underlying problem, 
but I am comfortable with it, my Asterisk server does not care, and it's
transparent to my users,so why not?
>Would you accept the need for a weekly reboot of your Nortel Option 11?
How 
>about if your car required you to remove the battery for two minutes once a
>month?  Your VCR?  How about a clock radio that had to be unplugged once a 
>week to fix weird little issues?
See above. Yes I would accept it, provided that it would not impact normal
operation and 
I as the administrator *or* enduser would not have to do anything special.
However, *this*
makes my sphincter shrink:

http://slashdot.org/articles/02/07/22/0615221.shtml?tid=126

Obviously, that's no good. Best comment:

RH support: Thanks for calling Red Hat! How may we help you?
Pilot: "Uhh.. I'm spiraling towards the earth, both my engines are out,
and
my display says 'kernel panic' in white text on a black
background."
RH Support: "And what is the system model?"
Pilot: "The F-22 jet.."
RH support: If you read linux-kernel-bugtraq, you will see that you should
have patched your kernel to 2.4.19-pre-alpha-revision-d before takeoff. But
no problem, this is Linux after all. Do you have another F22 on your LAN?
Just telnet in from there, su to root and restart sendmail.
Pilot: @#$*! Redhat! I'm switching to Debian if I survive!

>Hmm, I guess I won't be buying any Mitel equipment.  
Please don't. They are evil. Not so much the equipment but the corporate
philosophy is something
along the lines of "We're gonna hose our endusers more than any other
telecom manufacturer"

>Could you point to a specific issue? Any chance it could be backported
>to 1.0? This will mean less griff.
http://www.google.ca/search?q=tdm+static+site:lists.digium.com&hl=en&lr=&rls
=GGLD,GGLD:2004-23,GGLD:en&start=10&sa=N
>When do you reboot? every day?
Yes I do but the problem, for me, manifests itself every 1+N days so reboot
every day preempts this behavior since it *will* run fine for at least a
day. It's when I let it sit for 3 or 4 days that it happens. Zaptel reload
works too, but might as well reboot since it's just sitting there from 10 PM
to 6 AM. 

AFAIC, it's actually minor, we run 4 cordless phones that are used casually,
a few times a day so no biggie if the problem happens during business hours.
It does not affect any other operation of the server, SIP & IAX &
SpanDSP,
Sendmail, cdr, everything is 100%. This is the only fly in the ointment.

>Agreed but again -- you designed for this specific purpose.  Asterisk
isn't
>meant to boot up, answer the phone, process the call and shut down again 
>until the next ring.  (This would be an interesting approach to power
savings 
>though if your system boot time was fast enough and call volumes varied 
>enough to make it worthwhile.)
That's a friggin cool idea. I wonder if someone on the list will be inspired
enough to run with this:

http://www.linuxbios.org/index.php/FAQ#What_is_LinuxBIOS.3F

Current fastest boot time is 3s. You can have a flash filesystem for
Asterisk, which is supported. How would ring detection work? You would need
a "Wake on POTS" or something like that.

--On September 14, 2005 10:34:29 AM -0600 Colin Anderson 
<ColinA@landmarkmasterbuilder.com> wrote:
> Disclaimer: Not a troll
>
> I'm curious as to this obsession with uptime is. All of the posts of
this
> type are along the lines of "After X days, Y thing does not work but
if I
> reload or reboot, it's OK" - so why not cron a reboot? Is it
considered
> bad form or something like that? I reboot every night whether it is
> needed or not, not afraid to admit it, and everything works fine for me.
>
You obviously come from a land of Windows.  I have machines that are up for 
years at a time internally, and externally months, depending on how often 
security patches or other critical upgrades need applying.  That's how 
computers should run.  MS is the crap that gets people into this damned bad 
habit of oh i rebooted and it's fixed.  No it's NOT fixed, you just got
rid
of the symptoms, the BUG is still there.

For most UNIX and derivitaves/workalikes reboots are unnecessary.  Only if 
something is broken, or seriously wedged.
> We also do the "Sunday reboot" of all of our Windows servers as
well as
> restarting all of the critical services such as IIS , SQL, Exchange etc
> nightly. It helps, a lot (Exchange is a notorious memory leaker)
Yup, MS you need to because MS is broken.
> I've also heard it said, something along the lines of: "If you
have to
> reboot, your server isn't set up correctly" to which I say piffle.
Even
> NASA has rebooted the Mars probes after they land and I understand that
> they run VXWorks, incidentally, the same RTOS that my Mitel 3300 uses,
> and *even Mitel* recommends periodic reboots, which we duly cron every
> night, 2 AM.
Mitel isn't exactly known for telco grade operations.  And telco is usually 
the one that everyone quotes as being some of the most reliable systems in 
the world.  Imagine if Cisco or Juniper had this flawed mentality?  Or 
Fore?  Or IBM?  Or Acatel?  Or Nortel?  Yes I probably sound like a kook or 
a troll but so be it.

If a system is doing repeatable operations, and cleaning up correctly after 
itself, and doing periodic maintenance correctly, a reboot should never, 
under any circumstances be necessary.  Even restarting the services should 
be a rare item (hence one of my beefs with Apache mod_ssl).
> 24/7/365 installs aside, is there a reason why reboots seem to be frowned
> upon? Again, not trolling, just curious.