Historizing

[pfetrial]

Hello all,

I made some tests with a Python server . I found out that the minimal data rate change is around 10 ms .
Is it possible to start/stop historizing ? per example : to store 1000 data changes and then stop historizing then trigger/start again the historizing function to catch some interesting changes?
I tried to use the functions :
server.historize_node_data_change([myvar1,myvar2,mysin],period=0, count=100)
server.dehistorize_node_data_change([myvar1,myvar2,mysin])

but after the first stop/start historizing, the server makes an error..
Is there another way to historize on demand?
thanks

[AndreasHeine]

Hey,

why is it important to start and stop historizing?
keep in mind that a dynamic addressspaces are tricky for client applications because they dont know when something changes so they need to constantly check for changes this generates alot overhead on clientside aswell on the network.
in my oppinion its more usefull to constantly historize and if a trigger or event happens i read the history!

[pfetrial]

Hi Andreas,
Thanks for your reply.
It is important to stop the historization to avoid the overrule the storage with not important datas. Well, if you have a lot of memory/storage you can, as you pointed out, read the history when a special event occurs.
Concerning the problem of the call of start ans stop, It has nothing to do with the client because the error is generated by the server. Am i wrong?

[AndreasHeine]

in that case it might be easier with an Array[1000] and not opc ua historizing you can create a variant with one dimension, then you are free to do stuff you like and how you want!

by the way what is the errormessage you mentioned? could you post the stacktrace!

[pfetrial]

correct ! but we want to use the facilities of the historization of OPC ua ( display ) and In the case we use a Array of samples , we must write a specific software in the PLC .
Here is the stacktrace of the error :

Endpoints other than open requested but private key and certificate are not set.
Listening on 0.0.0.0:4840
Unknown message received FourByteNodeId(i=12208)
start histo call
stop histo call
start histo call
Error executing method call CallMethodRequest(ObjectId:{self.ObjectId}, MethodId:{self.MethodId}, InputArguments:{self.InputArguments}), an exception was raised:
Traceback (most recent call last):
File "C:\Python38\lib\site-packages\opcua\server\address_space.py", line 478, in _call
result = node.call(method.ObjectId, *method.InputArguments)
File "C:\Python38\lib\site-packages\opcua\common\methods.py", line 64, in wrapper
result = func(parent, *[arg.Value for arg in args])
File "C:\pfe\python_Filetransfer\minimalserverWithhistorizing.py", line 21, in methodstart
server.historize_node_data_change([myvar1,myvar2,mysin],period=0, count=2000)
File "C:\Python38\lib\site-packages\opcua\server\server.py", line 551, in historize_node_data_change
self.iserver.enable_history_data_change(node, period, count)
File "C:\Python38\lib\site-packages\opcua\server\internal_server.py", line 235, in enable_history_data_change
self.history_manager.historize_data_change(node, period, count)
File "C:\Python38\lib\site-packages\opcua\server\history.py", line 218, in historize_data_change
self.storage.new_historized_node(node.nodeid, period, count)
File "C:\Python38\lib\site-packages\opcua\server\history.py", line 91, in new_historized_node
raise UaNodeAlreadyHistorizedError(node_id)
opcua.server.history.UaNodeAlreadyHistorizedError: NumericNodeId(ns=2;i=2)

thx

[AndreasHeine]

opcua-asyncio/asyncua/server/internal_server.py

Line 233 in f189d87

async def enable_history_data_change(self, node, period=timedelta(days=7), count=0):

opcua-asyncio/asyncua/server/internal_server.py

Line 242 in f189d87

async def disable_history_data_change(self, node):

its part of the server.iserver
maybe to achive your target you might try that lowlevelmethods!

[oroulet]

Not being able to restart historyzing looks like a bug. Make a minimal example then make a bug report.
But concerning your use case, I really cannot imagine you really need such a feature. It really sounds like an architecture error, Ua is just a protocol, if you need to reduce data storing, it should not be done in protocol but your backend...

[AndreasHeine]

not a bug! stop method is for closing db connections before shutdown. maybe a missing api feature to pause the historization. what is missing is checking if the db file exist

[oroulet]

if we close the connection maybe we should be able to reconnect?
Anyway as written above I cannot see any use case that make sense. We should not add features that encourage bad practices

[AndreasHeine]

if we close the connection maybe we should be able to reconnect?

therefore we need to catch some sqlite errors if table already exist for example is missing (i just used our examples) atm we catch all errors ...

[pfetrial]

From my point of view, this use case is helpful , perhaps you don’t fully understand the use case : May I explain this in details: We have a machine controlled by a internal state machine. In the state “execute” ( in the meaning of ISA88) the machine is producing , in the others states the controlled items are not interesting . In the state execute ( during 180 secs ) it’s highly important to see how the machine is reacting with all the 60 items influencing in the process. This bunch of 60 vars is the image of the process and can be analyzed after the execution to optimize/debug the process . May this have a sense? In others stack , I have seen such Api to strop/restart histo.. Von: oroulet <[email protected]> Gesendet: Dienstag, 29. Dezember 2020 09:57 An: FreeOpcUa/opcua-asyncio <[email protected]> Cc: pfetrial <[email protected]>; Author <[email protected]> Betreff: Re: [FreeOpcUa/opcua-asyncio] Historizing (#375) if we close the connection maybe we should be able to reconnect? Anyway as written above I cannot see any use case that make sense. We should not add features that encourage bad practices — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#375 (reply in thread)> , or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAOAXGR7KVBGLE2TCUPEZJ3SXGKXBANCNFSM4VL2AD3Q> . <https://github.com/notifications/beacon/AAOAXGQB6SN7ZHUEB3TKF53SXGKXBA5CNFSM4VL2AD32YY3PNVWWK3TUL52HS4DFWFCGS43DOVZXG2LPNZBW63LNMVXHJKTDN5WW2ZLOORPWSZGOAAB5ATY.gif>

[oroulet]

OK. so you do not want the history server, but you want to stop historysing one or several nodes? I am still not sure about the use case but we should support that kind of things. Do you use the dehistorize() method? make a minimal code showing your issue

[pfetrial]

When you say «history server “ do you mean an external history server? And yes I want to stop historizing on several nodes. No I don’t use dehistorize() , as already written I use enable/disable history_data_change. Should I use something else? Von: oroulet <[email protected]> Gesendet: Dienstag, 29. Dezember 2020 11:09 An: FreeOpcUa/opcua-asyncio <[email protected]> Cc: pfetrial <[email protected]>; Author <[email protected]> Betreff: Re: [FreeOpcUa/opcua-asyncio] Historizing (#375) OK. so you do not want the history server, but you want to stop historysing one or several nodes? I am still not sure about the use case but we should support that kind of things. Do you use the dehistorize() method? make a minimal code showing your issue — You are receiving this because you authored the thread. Reply to this email directly, <#375 (comment)> view it on GitHub, or <https://github.com/notifications/unsubscribe-auth/AAOAXGU2RUGXDBIN4IQLFHTSXGTFFANCNFSM4VL2AD3Q> unsubscribe. <https://github.com/notifications/beacon/AAOAXGVHKHHWMCDE4RI54RTSXGTFFA5CNFSM4VL2AD32YY3PNVWWK3TUL52HS4DFWFCGS43DOVZXG2LPNZBW63LNMVXHJKTDN5WW2ZLOORPWSZGOAAB5BRA.gif>

[pfetrial]

So internally the server use sqllite to store the history ? Von: Andreas Heine <[email protected]> Gesendet: Dienstag, 29. Dezember 2020 11:27 An: FreeOpcUa/opcua-asyncio <[email protected]> Cc: pfetrial <[email protected]>; Author <[email protected]> Betreff: Re: [FreeOpcUa/opcua-asyncio] Historizing (#375) therefore we need to catch some sqlite errors if table already exist for example is missing — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#375 (reply in thread)> , or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAOAXGR44JUTSJORLHKIYZ3SXGVHZANCNFSM4VL2AD3Q> . <https://github.com/notifications/beacon/AAOAXGSNUZOAXXH22EGVBNDSXGVHZA5CNFSM4VL2AD32YY3PNVWWK3TUL52HS4DFWFCGS43DOVZXG2LPNZBW63LNMVXHJKTDN5WW2ZLOORPWSZGOAAB5BZI.gif>

[AndreasHeine]

if you use the default historymanager then yes

[pfetrial]

Yes I do. But does it use files for storage ? or does it use ram? Von: Andreas Heine <[email protected]> Gesendet: Dienstag, 29. Dezember 2020 15:47 An: FreeOpcUa/opcua-asyncio <[email protected]> Cc: pfetrial <[email protected]>; Author <[email protected]> Betreff: Re: [FreeOpcUa/opcua-asyncio] Historizing (#375) if you use the default historymanager then yes — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#375 (reply in thread)> , or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAOAXGSIUKSMF67ROOKU32TSXHTUZANCNFSM4VL2AD3Q> . <https://github.com/notifications/beacon/AAOAXGS5H23DKSSAP7EXRATSXHTUZA5CNFSM4VL2AD32YY3PNVWWK3TUL52HS4DFWFCGS43DOVZXG2LPNZBW63LNMVXHJKTDN5WW2ZLOORPWSZGOAAB5FNI.gif>

[AndreasHeine]

with sqlite is both possible depends on your code!

file:
server.iserver.history_manager.set_storage(HistorySQLite("my_datavalue_history.sql"))
ram:
server.iserver.history_manager.set_storage(HistorySQLite(":memory:"))

there is also a HistoryDict!