Welcome, Guest. Please login or register.

Author Topic: NeDi 1.0.7 - target availability to zero after some server crash  (Read 1989 times)

pc_sg

  • Sr. Member
  • ****
  • Posts: 271
    • View Profile
As in subject, one week ago my NeDi server crashed (more times...)
So I've upgraded it (was Debian 5.0 Lenny, now is a Debian 6.0 Squeezy).

All seems ok, but I've seen that the target availability graphc remains at zero.

Database corrupted?
There is something I can do to solve this "issue"?

Is not mandatory... I'm going to reinstall NeDi from scratch into some completely new (virtual) servers (one for each site, awaiting the new 1.0.8 or a much smarter new version of NeDi)

TIA

Paolo


rickli

  • Administrator
  • Hero Member
  • *****
  • Posts: 2711
    • View Profile
    • NeDi
Re: NeDi 1.0.7 - target availability to zero after some server crash
« Reply #1 on: April 20, 2012, 06:25:11 PM »
Have a look at the last TRRD: lines in the discovery log. The Mon= part draws the availability graph. If you have numbers there, you may have a corrupted top.rrd. Have you tried rrdtool dump top.rrd?
Please consider Other-Invoices on your NeDi installation for an annual contribution, tx!
-Remo

pc_sg

  • Sr. Member
  • ****
  • Posts: 271
    • View Profile
Re: NeDi 1.0.7 - target availability to zero after some server crash
« Reply #2 on: April 23, 2012, 11:41:01 AM »
Last lines (two rows) from

"rrdtool dump /var/nedi/rrd/top.rrd"

Code: [Select]
<!-- 2012-04-23 07:30:00 CEST / 1335159000 --> <row><v>3.9278293132e+01</v><v>6.0392933908e+01</v><v>3.2973407959e-02</v><v>0.0000000000e+00</v><v>2.8172787625e+02</v><v>0.0000000000e+00</v><v>0.0000000000e+00</v><v>5.1542847979e+03</v><v>7.6147152021e+03</v><v>4.8000000000e+01</v><v>0.0000000000e+00</v><v>0.0000000000e+00</v><v>0.0000000000e+00</v><v>0.0000000000e+00</v><v>2.1129743452e+01</v><v>1.5399938316e+01</v><v>3.3615402036e+01</v><v>0.0000000000e+00</v></row>
                        <!-- 2012-04-23 10:00:00 CEST / 1335168000 --> <row><v>9.3312705610e+01</v><v>1.1157474674e+02</v><v>9.6276679594e-01</v><v>1.7690220312e-01</v><v>3.5039197587e+02</v><v>8.4161894600e-01</v><v>0.0000000000e+00</v><v>5.6073513240e+03</v><v>7.1452713706e+03</v><v>4.8000000000e+01</v><v>0.0000000000e+00</v><v>0.0000000000e+00</v><v>0.0000000000e+00</v><v>0.0000000000e+00</v><v>2.8000934264e+01</v><v>1.7367638585e+01</v><v>3.3150735823e+01</v><v>0.0000000000e+00</v></row>
                </database>
        </rra>
</rrd>

Should I add any debug command in my scheduled discovery?

TIA

Paolo

Edit: no TRRD: lines in verbose output (this means your "log" ?). Why?
« Last Edit: April 23, 2012, 04:07:58 PM by pc_sg »

rickli

  • Administrator
  • Hero Member
  • *****
  • Posts: 2711
    • View Profile
    • NeDi
Re: NeDi 1.0.7 - target availability to zero after some server crash
« Reply #3 on: April 23, 2012, 07:34:17 PM »
In System-Services is there always a thread running? Could be a hanging one and can be reset with the red X (best done, if there's no discovery running)
Please consider Other-Invoices on your NeDi installation for an annual contribution, tx!
-Remo

pc_sg

  • Sr. Member
  • ****
  • Posts: 271
    • View Profile
Re: NeDi 1.0.7 - target availability to zero after some server crash
« Reply #4 on: April 24, 2012, 07:54:59 AM »
I've just rebooted server.
I'll verify situation later. Hope I understand correctly real task situation.

I'm using Steffen way to launch parallel discovery, so concurrent tasks may happen.
I know that a real multithread handling is better  ;)

Thanks

Paolo

rickli

  • Administrator
  • Hero Member
  • *****
  • Posts: 2711
    • View Profile
    • NeDi
Re: NeDi 1.0.7 - target availability to zero after some server crash
« Reply #5 on: April 25, 2012, 06:02:32 PM »
It would be, but also tricky to implement properly! I think it's not feasable for me to do on OBSD in addition...

Look on the bright side, manually starting discoveries keeps you in control of the bandwidth used for a discovery by not having 5 devices queried over the same WAN link for example...
Please consider Other-Invoices on your NeDi installation for an annual contribution, tx!
-Remo

pc_sg

  • Sr. Member
  • ****
  • Posts: 271
    • View Profile
Re: NeDi 1.0.7 - target availability to zero after some server crash
« Reply #6 on: April 26, 2012, 08:11:43 AM »
TNX Remo!

After reboot availability seems ok.

About OBSD, why don't you migrate on Debian, that is really stable ?
I know there is a lot of other distributions (no flame, please! I'm not a DebianBoy!) , I've choosed Debian because of a lot of people that I trust recommend it for a clean, stable, easy and really supported Linux Server distribution.

Or OBSD is better for you due to easily scriptable installation?

(only curiosity)

Paolo