Add snmp inventory Switch Huawei

[xfilho]

Bug reporting acknowledgment

Yes, I read it

Professional support

None

Describe the bug

I'm trying to add the huawei switch model S5720-52X-PWR-SI-AC in glpi using netdiscovery/netinventory, but I'm not succeeding. The agent probably does not have the device's SNMP identifier in its internal database. I added to the sysobject.ids file "2011.2.23.335 Huawei NETWORKING S5720-52X-PWR-SI-AC" as per documentation, but that didn't work either.

Error:

[debug] [thread 1] #1, full match for sysobjectID .1.3.6.1.4.1.2011.2.23.335 in database
[warning] Aborting netdiscovery task as it reached expiration time
[error] 1 devices scan result missed
[error] job 1 aborted
[debug] All netdiscovery threads terminated

To reproduce

1. On the glpi server, where the agent is installed, I executed the command: glpi-netdiscovery --credentials version:2c,community:********* --debug --host IP-SWITCH

Expected behavior

I can't inventory the huawei switch S5720-52X-PWR-SI-AC

Operating system

Linux

GLPI Agent version

1.4

GLPI version

10.0.x (See additional context below)

GLPIInventory plugin or FusionInventory for GLPI plugin version

No response

Additional context

No response

[g-bougard]

Please people, stop open a bug issue if you don't provide evidence of a bug. First open a Q&A discussion.

@xfilho please provide the full output in debug2 with:

glpi-netdiscovery --credentials version:2c,community:********* --debug --debug --host IP-SWITCH

When you see a it reached expiration time warning, you should also say in how many time that expiration was reached... but probably the full requested output will tell it.

Also, you should firstly not provide the output after a sysobject.ids update or you are no more complaining about the same product. If you see a partial match message (and you should) when sysobject.ids is not modified, this is also normal. The only change you're doing when you modify sysobject.ids for such device is to set the MODEL if it's not set.

[g-bougard]

Hi @tech62
as you seem to point out, the problem is the set expiration, not the minimal one. I guess this would have worked if you just have changed the previous line to:

my $target_expiration = 18000;

Anyway, I'm updating the code to make the target expiration a function of backend-collect-timeout configuration, actually only used for inventory task.

[tech62]

Hi @g-bougard !
Thanks for the fix !
The informations are not correct for some ports, how can i use specific mibs with glpi-netinventory from windows ?

[g-bougard]

@tech62 can you explain what is not correct for these ports ?

For you to know, I just pushed 64e405c to make expiration a function of backend-collect-timeout. The commit also add the option to glpi-netdiscovery & glpi-netinventory scripts. To not change the default which works for most people, for netinventory task, I just multiplied the value by 5 (180 by default means 900s, so 15 minutes). So in your case, you'll have to set it to 3600 to have 18000 as resulting expiration.

Next nightly build will include the update.

[tech62]

Nice!,
Wrong vlans returned per example

[g-bougard]

To handle such problem, you'll have to provide a full snmpwalk output with right format and list everything wrong in your point of view: gotten value vs expected one.