Minor stutus of nodes and error message in object details in new version 2.1

Started by CrUser, August 30, 2017, 04:38:33 PM

Previous topic - Next topic

CrUser

Dears,

I upgraded NeteXMS server from 2.0.8 to 2.1 version but in new version I got same strange notification and error in work of server.
First, at a few cisco switch server can't show port view and list of interface. After I change tab on Object details server generate message box with error
"An Error has occurred when activation this view" (attach). I have never seen this message before in the old version of server.
Second problem is status of some nodes. For a example, after one of interface goes down node change status to critical. When this interface go to normal state
node go to some strange "Minor" state. I have checked all alarms of this node and interface status and all is in the normal state but icon of node is still o "Minor" state. When this occurred I can only recreate this node (delete and create new node) to change his status to normal.

Also, I have question how I can change events severity level for interface down of some node from "Critical" to "Warning" or "Minor".

If anybody have some idea to resolve these problem please help me.

best regards,
Tomislav

Tursiops

Regarding the node status, you may not be seeing all alerts after the migration to 2.1.
Ensure your user has "View all alarm categories" permissions. This is new to 2.1 and is not granted automatically to existing users (including admins).

You can change the status calculation per interface. Select the interface, open properties and Status Calculation.
When an interface is not in the "expected" state, it will generally trigger a "Critical" status. If you select "Severity based" under "Propagate status as" and switch Critical -> Critical to Critical -> Warning, it will only trigger a warning status on the node (for that interface only).
Alternatively, if there are interfaces you don't care about because they are not "always on", just set the expected state to "Ignore".

Ēriks Jenkēvics

The "View all alarm categories" permission should automatically be added to the Everyone user group after performing the DB upgrade when upgrading to 2.1

CrUser

Thanks a lot,

I have changed "View all alarm categories" permission for admin and now I can see all alarms for any node and after resolve all alarms node goes to normal state.
Also, I have changes propagation status as Critical -Warning and now I get warning alert for port down on switch.
Now I have only one problem with interface and port view on object details tabs, for some switch I can see interface table and ports but
for other ports tab is empty and interface tab is not refresh with list of interface. Problem usually with stacked switch but find same problem with normal switch. If you have some advice how I can resolve this bugs please help me.

Have you experience with new future Outstanding alarm remainder. I have set one sound for this option but I when alarm occured noting hapened.?
Is there any server parameter to set time of outstanding reminder?


best regards,
Tomislav



Tursiops

Eriks:
That didn't seem to happen during the upgrade at our end. Considering the number of post-2.1 upgrade forum posts which are fixed by checking the "View all alarm categories" permissions, I have to assume something may not always work as expected during that DB upgrade?

Tomislav:
If a switch is showing anything in the Ports tab depends on if NetXMS has a driver for the device in question. For us that tab works with HP and Cisco gear, but not with Brocade, Netgear and TP-Link "Smart" switches.
For the Interface tab, select the properties of the node, select Polling and test the Enable and Disable options under "Use ifXTable for interface polling", run a Configuration Poll after each and see if the Interfaces tab populates. If it doesn't, the switch in question may have some vendor specific way of providing interface information (which would require a specific driver for that hardware).
I am actually not sure about the Outstanding Reminder. I tend to not leave my alerts in "Outstanding" state, but Acknowledge them.
For testing purposes I just enabled the "Show pop-up reminder for outstanding alarms", set a sound and triggered some events which I didn't acknowledge. Eventually I got a pop up with a list of outstanding alarms together with the sound. I didn't time it, but seemed to happen/repeat every five minutes?

Ēriks Jenkēvics

Hmm... That`s very interesting. I just tested the DB upgrade from 2.0 to 2.1 and it worked for me, it also worked when initialising a new DB. Could you please verify that the group "Everyone" has the "View all alarms" access right set? And if it changes anything if you disable it (or enable it)?

Tursiops

If the "Everyone" group is the only one getting that flag set during the upgrade, then maybe that's the issue. We are certainly using more specific groups at our end, hence the need for us to set this manually.
Considering that pre-2.1 every group effectively had this permission (as no categories existed and everyone could see everything), maybe all user groups should have it enabled by default upon an upgrade, not just "Everyone"?

But this is kind of hijacking the thread now. :)

CrUser

Thanks for help,

Maybe I didn't describe problem with port and interface show enough.
On the object browser tab a can see all interface of every switches and on the object details tab I can see overview, alarms, last value and performace as well.
But when I select Interface or Ports tab on Object Details for some switches interface table and ports is not show (or stay on last switch port list). Also, when I select tab overview after selection of ports or interface I got error. I sent you error picture in the attachment.

I have check information data in Overview tab for all switches and I can see anything suspected or wrong.
In my system I use cisco SF500 and SF500 stacked switches.


best regards,
Tomislav




Victor Kirhenshtein

Hi,

it's a bug in UI, we will fix it in next patch release.

Best regards,
Victor

CrUser

Hi Victor,

Thank you form you replay.
I hope that will be resolved with next patch.

best regards,
Tomislav

Ēriks Jenkēvics

CrUser and Turslops, have you made any changes to the Everyone group? If not, could you, please, verify that it exists in your DB? You can run this query to find out:
" SELECT * FROM user_groups WHERE name='Everyone' "

Regards,
Eriks

Tursiops

Hi Eriks,

We have an Everyone group, but it doesn't contain any actual users (at least it's not listing any).
Note that we are using LDAP. Not sure if Everyone only applies to Local accounts (though the few that exist aren't showing as members either).

Cheers

Ēriks Jenkēvics

The Everyone group includes all users, even if there are not any listed, it`s sort of like the built in System user which has all privileges. Anyway, the issue should be solved once and for all in the new release :).

Regards,
Eriks