fortinet_get_system_ha_status.textfsm parse error for non "OK" HA Health Status #1859
Comments
|
@pnpestov |
|
@pnpestov The parser is failing in the first |
Starting from the 6.4 branch, there is incorrect output in the tests. You can look at the 6.2 branch, everything is correct there. |
Unfortunately, all the devices are in operation and I can't pull the cable out of the lan1 port. But you can just put four spaces each, similar to my last output when executing the command. I think it will be like this |
No, I don't think so. If only the carriage translation character is '\n' |
|
@pnpestov So far I haven't detected an issue other than the leading white space on the If you'd like, try the changes I've made to my feature branch for this. |
|
@mjbear |
|
@mjbear Receive |
Are you indicating you reordered those two lines? Ah,
Which is likely why you're mentioning void (or null). ➡️ Edit: What's wild though is that the existing test data functions fine (no changes to their yaml test output). |
Yes |
|
Your recent changes result in the same error, but the content of the line is as follows: "Rule Line: 37. Input Line: HA Health Status: ." |
Ugh. Edit: |
Not quite like that. After all, the OK state is still processed correctly. But if the status is not OK, then the error presented above appears |
|
@pnpestov I decided to change a few items (so best copy the entire template to be safe).
Please give this a shot. 🤔 There has to be some sort of white space I'm missing within this GH issue thread. Edit: Gah, this has issues too. Hmmmf |
|
I'll try it now. Apparently, the variable ${HA_HEALTH} is overwritten several times |
|
Your last template for the OK status works correctly, and for the WARNING status it does not return anything in the desired variable, but the error does not occur. |
|
@mjbear I get the following values of the HA_HEALTH variable at the output |
@pnpestov
The So far our test data shows we have either
|
|
@pnpestov |
It seems to me that the line |
I checked, there is no error, but the variable also outputs [] in the case of WARNING |
Depends on which capture group is in use (please read my note below).
The situation you describe is why I made the capture group require at least one non-white space character and then the
I don't expect any change in behavior even if we were to make the non-capturing group |
I get the same [] at the output |
|
@pnpestov So far the template changes I made were ran against the test data you provided as well as the existing test data. Those tests parsed successfully. I haven't seen what you have configured/coded so I have an unfounded suspicion that somehow another copy of ntc-templates is being accidentally referenced. Hopefully you can humor me by setting up a basic script (see below 👇) that should tell us for sure if it's a template syntax issue. (That is - for testing right now don't have Netmiko do the parsing via 🙏 I hope this helps identify the issue. 🙏 Netmiko to retrieve text, but manually parse with TextFSMBelow you'll find an untested script (don't have Fortinet gear) modified from this example. import textfsm
from netmiko import ConnectHandler
device = {
'device_type': 'fortinet',
'host': 'X.X.X.X',
'username': 'test',
'password': 'password',
# 'port' : 22, # optional, defaults to 22
# 'secret': 'secret', # optional, defaults to ''
}
net_connect = ConnectHandler(**device)
output = net_connect.send_command("get system ha status")
template = open('/tmp/fortinet_get_system_ha_status.textfsm')
re_table = textfsm.TextFSM(template) # initialize textfsm object
fsm_results = re_table.ParseText(output) # parse output text with textfsm object.
template.close()
"""
ParseText output returns back a list of tuples.
First tuple is the header, every subsequent tuple is a row.
Let us make a dict with relevant key value pairs out of this!
"""
results = list()
for item in fsm_results:
results.append(dict(zip(re_table.header, item)))
print(results)Parsing with content in local filesAnd if you'd like to parse the raw output against the template, update the paths to reflect your local system and give it a try. (json.dumps solution for prettier output.) import json
# from pprint import pprint
import textfsm
with open('/tmp/fortinet_example.txt') as fh:
raw = fh.read()
tpl_fh = open('/tmp/fortinet_get_system_ha_status.textfsm')
tfsm_parser = textfsm.TextFSM(tpl_fh)
tpl_fh.close()
parsed = tfsm_parser.ParseText(raw)
# pprint(parsed)
print(json.dumps(parsed, indent=4)) |
|
@mjbear Netmiko HA Health Status: From file HA Health Status: Software versions: |
|
@pnpestov
I suspect this indicates there's some character that apparently isn't white space at the end of the "HA Health Status:" line when Fortinet is in an unhealthy state.
This is good in that it shows us the template is functional on text in a file (where some terminal chars may have been lost) and points to there being extra characters. Instead of a Note: If the regex on the previous line doesn't work you can try making that above rule look like ✨ I've updated that feature branch on my fork of ntc-templates if you're feeling in the mood for another test. |
Yes, there is a desire! Moreover, the ideas have not ended yet and at least this ticket improves the template. |
ISSUE TYPE
TEMPLATE USING
SAMPLE COMMAND OUTPUT
SUMMARY
Version: FortiGate-40F v7.0.15,build0632,240401 (GA.M)
STEPS TO REPRODUCE
Reproduce non "OK" HA Health Status in two lines. For example, disable the lan1 (HA Monitor Interface) or lan2 (HA Heartbeat Interface) work link on the slave node.
EXPECTED RESULTS
Get the current value of HA Health Status
parsed_sample:
"WARNING: FGT40FYYYYYYYYYY has mondev down"
or
"WARNING: FGT40XXXXXXXXXXX has hbdev down"
"WARNING: FGT40YYYYYYYYYYY has hbdev down"
or other options in combination
and continue executing the script
ACTUAL RESULTS
The text was updated successfully, but these errors were encountered: