Details
Major Description
Attached changes add support for the FMT04-CH1U optical device, which can accommodate a number of optical modules (this patch adds support for fibre amplifier and an optical-electrical-optical convertor). The MIBs and the SNMP agent supplied with the device require a layer of index translation and use of get instead of walk.
I have tested the two attached patches (one for CE-17.9.0 and the other for 18.1.9002) and confirmed they both work as expected. You can see the sensors processed by poller.php in an attached file. Let me know if you need any additional information to merge these changes.
Attachments
Activity
The PSU declarations in changeset 9006, file mibs.inc.php need some fixing. There are two power supplies and each has a state:
OAP-NMU::power1State.0 = INTEGER: on(1)
OAP-NMU::power2State.0 = INTEGER: on(1){{}}
{{}}
Thanks for this Mike, we pulled the patch to our production copy overnight and it has found (almost) everything correctly. For some reason, none of the units are showing PSU 1 (they all show their PSU 2 correctly). Apart from this it has found everything OK. Denis is going to look at it now to try to see why. Thanks again!
JFYI, snmpwalk will emit a warning for the EDFA MIBs because of malformed module revision. Also OEO object index 13 (Tx power threshold) is present in the MIBs only, the agent does not have it. And walk will always get the same tree regardless of what was the request. There are reasons this patch is a bit complicated but of course feel free to improve it.
infrastation yes, I see (MIBs).. 
anyway some parts were made incorrect:
- if you not use valid NamdeOid, than use numeric oid (only), ie: enterprises.40989.10.16.{$slot_no} -> .1.3.6.1.4.1.40989.10.16.{$slot_no}
- detect os should use definition detect if possible (that possible in 99% cases, also in this)
- hardware/version/serial in your case anyway use only OAP-NMU mib, that shold be in os/mib definitions
- MIB files needed anyway and should be stored in subdir (fscom)
I will rewrite this parts from your patch 
Please see the comments in fsfmt-pseudo-mib.inc.php, FSFMT-PSEUDO-MIB does not actually exist, it is a virtual wrapper around the original MIBs, which are not very workable as they are (see attached archive). Also this SNMP agent supports version 1 only, it does not support walk properly (snmpwalk rejects its response without -Cc), and the only objects it ever supports are the ones in original MIBs plus the following:
SNMPv2-MIB::sysDescr.0 = STRING: GLCY SNMP
{{ SNMPv2-MIB::sysObjectID.0 = OID: SNMPv2-MIB::sysObjectID}}
{{ DISMAN-EVENT-MIB::sysUpTimeInstance = Timeticks: (51859500) 6 days, 0:03:15.00}}
{{ SNMPv2-MIB::sysContact.0 = STRING: }}
{{ SNMPv2-MIB::sysName.0 = STRING: }}
{{ SNMPv2-MIB::sysLocation.0 = STRING: }}
So perhaps it would make sense to suppress most of the other MIBs during discovery because OSPF currently "detects" area 0 on that device (which does not do OSPF). I would do it if I knew the right way to do. Let me know if you need any additional information.
1. I not see FSFMT-PSEUDO-MIB mib file
2. yah, this is very old style of adding os support (need some rewriting to definitions)
3. better if give access to example device by snmp.. if possible, write me to mike@observium.org (my dev ip for ccess 77.222.50.30)
Just to ease the deployment of this, as with the previous patch we published with Adam for the SmartOptics DWDM chassis under http://jira.observium.org/browse/OBS-1089 - we can if required provide SNMP forwarding to one of these chassis for your viewing pleasure...
However, for what it's worth, i can confirm that the output is good from the end-user's perspective and Denis has done a tidy job of this
Confirmed - all good now thanks!