Мониторинг Windows Update

[HaronSSG]

скрипт отрабатывает некорректно.
при запуске скрипта выдает 0
а реально

http://smotr.im/efJL

[westwind]

Hi guys,
Can somebody post the solution in English please ?
I'm suffering from the same problem of windows update checks.
It looks like that the process takes longer time than 30 sec and then reports Not_Supported.
please post a correct script to fix that or an alternative solution.

thanks,
Westwind

[Mox]

Hi guys,
Can somebody post the solution in English please ?
I'm suffering from the same problem of windows update checks.
It looks like that the process takes longer time than 30 sec and then reports Not_Supported.
please post a correct script to fix that or an alternative solution.

thanks,
Westwind

Hi!
One of the obvious idea is to make check active. I.e. run periodically your script on windows server using the scheduler. And at the end of the script use zabbix_sender to send output to the zabbix server.

[westwind]

Hi!
One of the obvious idea is to make check active. I.e. run periodically your script on windows server using the scheduler. And at the end of the script use zabbix_sender to send output to the zabbix server.

Mox, Thank you for the quick response !

Do you mean that the session of getting the list of items from Z server should be skipped ?

As far as know these two sessions have to be done:
1- the agent does an active checks and requests the list of items assigned to the host in order to know what to be monitored. then the server responses and sends the the items (hostname, a key, and a delay per item)
2- the agent disconnects and starts to collect data and periadically sends them back to the server (data included: hostname, key, value, and clock per item). then the resver will response with a summary.

You've mentoined:
- active win.update checks through running a scheduler script periadically on the host.
- using zabbix_sender (at the end of the script) to send the collected data to z server. right ?

Where to put your steps there ? somewhere between ? or after/before ? or should one of my steps be neglected ?

westwind

[Mox]

Mox, Thank you for the quick response !

Do you mean that the session of getting the list of items from Z server should be skipped ?

As far as know these two sessions have to be done:
1- the agent does an active checks and requests the list of items assigned to the host in order to know what to be monitored. then the server responses and sends the the items (hostname, a key, and a delay per item)
2- the agent disconnects and starts to collect data and periadically sends them back to the server (data included: hostname, key, value, and clock per item). then the resver will response with a summary.

You've mentoined:
- active win.update checks through running a scheduler script periadically on the host.
- using zabbix_sender (at the end of the script) to send the collected data to z server. right ?

Where to put your steps there ? somewhere between ? or after/before ? or should one of my steps be neglected ?

westwind

i'm sorry, my fault (
I meant switch item to zabbix traper, not active check. it's always mixed in my head, sorry

[westwind]

i'm sorry, my fault (
I meant switch item to zabbix traper, not active check. it's always mixed in my head, sorry

No problem. It's fine!

But this solution is almost impossible, because I have more than 500 hosts (windows servers) monitored by zabbix. A zabbix traper should then be initiated per host !

Is there really no way to increase the interval/delay/clock/time-stamp whatever it's called at the agent side or at the server side ?

[Mox]

No problem. It's fine!

But this solution is almost impossible, because I have more than 500 hosts (windows servers) monitored by zabbix. A zabbix traper should then be initiated per host !

Is there really no way to increase the interval/delay/clock/time-stamp whatever it's called at the agent side or at the server side ?

I didn't understand your statement about trapper.
I hope you have template for your 500 win hosts, and item that runs your script is in this template. Open this item in template and see at 'Type' field, this is drop-down list where you can change it to 'Zabbix trapper'.
There is another zabbix process which is called trapper. It collects alike items.

[westwind]

I didn't understand your statement about trapper.
I hope you have template for your 500 win hosts, and item that runs your script is in this template. Open this item in template and see at 'Type' field, this is drop-down list where you can change it to 'Zabbix trapper'.
There is another zabbix process which is called trapper. It collects alike items.

Definitely I understood where you were talking about.So I just need to change the Win.Update.Check item in a template which is assigned to the 500 hosts. please see these screen-shots:




And what is next ? Should I make more change on this ?

[Navoyenok]

Mox, доброго времени суток.

Если я не ошибаюсь, то для того, чтобы элемент данных типа "Zabbix trapper" чего-то получил, необходимо, чтобы что-то в этот элемент данных отправило данные. В данном случае, этим что-то является скрипт на стороне Windows-сервера, который должен запускаться по рассписанию, определять количество не установленных Update-ов и отправлят Zabbix-Sender-ом данные. Я думаю, что именно это "смущает" нашего анлоязычного гостя.

Может быть это не так, но именно настройка шедулеров на каждом подверженному мониторингу Windows-сервере и меня останавливает от внедрения мониторинга данного вида событий.

С уважением, Navoyenok

[westwind]

Yes exactly. because if I change Zabbix Agent (Active) to Zabbix Trapper, It will going to do PASSIVE CHECKS.. no more ACTIVE CHECKS. Am I right ?

[Mox]

Thanks, Navoyenok!
Westwind, I've understood now. You don't want to set scheduler task on all 500 hosts, right? Yes, it's a problem.
And yes, second screenshot is what i'm talking about + zabbix_sender at the end of script.
If 'Allowed hosts' is not set, then any host can send data for this item I think. Try without 'allowed hosts' and if it works try to set it to something like {HOST.HOST}, {HOST.CONN} or similar.

Maybe, if all your win servers are in domain, is there any ability to make global policy or global scheduler task for all hosts? (I'm not win or AD master, sorry)

[Mox]

Thanks, Navoyenok!
Westwind, I've understood now. You don't want to set scheduler task on all 500 hosts, right? Yes, it's a problem.
And yes, second screenshot is what i'm talking about + zabbix_sender at the end of script.
If 'Allowed hosts' is not set, then any host can send data for this item I think. Try without 'allowed hosts' and if it works try to set it to something like {HOST.HOST} or similar.

Maybe, if all your win servers are in domain, is there any ability to make global policy or global scheduler task for all hosts? (I'm not win or AD master, sorry)

Here is explanation for 2.0 https://www.zabbix.com/documentation...owed&s[]=hosts

and for 2.2

https://www.zabbix.com/documentation/2.2/manual/config/items/itemtypes/trapper?s[]=allowed&s[]=hosts

As i can see user-MACRO in 'Allowed hosts' will work only in 2.2, but I would try MACRO like {HOST.HOST} or {HOST.CONN} because it's not USER macro.

[westwind]

unfortunately they are not (in the same domain). however I'll ask our Microsoft person tomorrow.

Actually it's all about the timeout value:



And the real update process takes (sometimes) longer than 30s.

[westwind]

Hey wait a moment ! I noticed something interesting !

1- Update Interval (in sec) -from the 1st snapshot
2- Refresh Active Checks
3- Timeout

May be we get it worked if we can tune (synchronize) these values with each others ?

any idea ?

[Mox]

Hey wait a moment ! I noticed something interesting !

1- Update Interval (in sec) -from the 1st snapshot
2- Refresh Active Checks
3- Timeout

May be we get it worked if we can tune (synchronize) these values with each others ?

any idea ?

These values mean different things.
Just in case, when you observed that script takes more than 30 sec, did you tried it manually (run on objective host or used zabbix_get) or through zabbix server?
Have you set maximum timeout (30) on agent side and server side?