v1.2.0+ Enhancements and Known Bugs

[PackeTsar]

Logging enhancement requests and known bugs here. If you have reported a bug or feature request and it is accepted as such, then it will show up here and your original issue will be closed.

[PackeTsar]

Enhancements

• Enhancement 1: Add a global keystore setting to allow injection of variables into all keystores (Global Variables #64)
• Enhancement 2: Add caching options into ZTP config to avoid need to restart the service after changes (Request // re-cache templates or keystores without 'service restart' #34)
• Enhancement 3: Add the ztp request lookup feature (Request // Command to query external-keystore; ztp query <keystore_id> #28)
• Enhancement 4: Add option to ztp request merge-test <keystore_name> (allow-logging) to write the config file out to the log file described by the merged-config-to-custom-file setting
• Enhancement 5: Add ztp clear provisioning <keystoreid> option to clear the provisioning info for a specific keystore
• Enhancement 6: Build external script functionality, allowing FreeZTP to kick off an external job after a certain action.

[PackeTsar]

Bugs

• Bug 1: (FIXED) Rasperian GUI broken after FreeZTP install (GUI login issues after installing V1.1.0 FreeZTP on a RaspberryPI #41)
• Bug 2: (FIXED) show downloads live concurrent downloads cycles through downloads on one line (dev1.1.0m // Show Downloads output for concurrent downloads of the same file #35)
• Bug 3: (FIXED) Malformed CSVs will cause a crash. Need to fail gracefully (dev1.1.0g // Bad csv format breaks 'ztp' commands #26)
• Bug 4: (FIXED) Possible issue with the persistent store. Need to reproduce to confirm. (Persistent store error #22)
• Bug 5: (FIXED) show downloads and show provisioning do not sort by time (some issues after OS update #47)
• Bug 6: (MITIGATED) /etc/ztp/pdb tends to get corrupted on Pi's. Need to move onto TinyDB
• Bug 7: (FIXED) If an IDArray exists, but its referenced keystore does not, ZTP will throw an exception (IDArray pointing to non-existent keystore fails ZTP #56)
• Bug 8: (FIXED) RHEL seems unsupported (Support RHEL7 #59)
• Bug 9: (FIXED) show downloads live limit output to 20 lines (dev1.1.0m // Show downloads live have a max line limit? #36)

[PackeTsar]

From v1.1.0

Enhancement ideas carried from v1.1.0 tracker

• Slack notifications?
• Google Sheets integration
• Easy config backup and restore
• Add show config <section> commands
• Push .py script to Cisco IOS to automate configuration?

[PackeTsar]

Completed

Enhancements and Bugs included in 1.X releases

• Bug: DHCPD settings autocomplete fails (DHCPD Scope Autocomplete Fails #50) [v1.2.0a]
• Enhancement: Add ztp service dhcpd (start|stop|restart) command-set [v1.2.0b]
• Bug: SNMP discovered information is now included in template merge data under snmpinfo.<oid_object_name> (Ability to reference the SNMPOID variable names from within the templates? #49)
  • Also had to include a fix to modify any SNMP OID objects with a hyphen in the name to replace with an underscore. Fix automatically runs during the upgrade procedure.

[EarlRamirez]

@PackeTsar If you need any data regarding #22 issue I have loads of it, I believe you are correct with the concurrent writes. I see this every time I deploy a few hundred switches. Even after deleting the pdb and purging the logs the error still persist. I have also seen where the temporary ID ZTP- is given to the switch hostname. This is very common during large batches of switches, for example, fifty. Now I power on ~20 to try and get past that.

[kbarton1212]

@PackeTsar @EarlRamirez I am running into this same issue on a Rpi 3b. I am running Raspbian Buster, and I need to manually issue the following commands in order for the ztp service to start and run, because by default after the RPi boots, it shows as failed with (code=exited, status=1/FAILURE):

mv /etc/ztp/pdb /etc/ztp/pdb.bak
ztp service restart

I was messing around with the autostart on the RPi, and pasting those commands in there in an attempt to get it up and running at boot, but no luck. Any chance you've had any luck @EarlRamirez ?

[EarlRamirez]

@kbarton1212 - what I observed is concurrent writes corrupt the file and deleting/moving the PDB file works. Initially, I would truncate the file, this removes the provisioning and tracking JSON collections, and since the file already exists it will not be created. Therefore deleting/moving is the workaround.

The expert @PackeTsar can shed some light on this, but what I have been working on is to try and have both provisioning and tracking write to MongoDB which should resolve the issue with the ZTP service failing because of PDB being corrupted.