Cloud functions synced multiple times

Technical Issue Reporting
#1

Using this script to save our cloud functions, we noticed that the aftersave functions were saved in the cloud function Moralis UI more than once.

moralis-admin-cli watch-cloud-folder --moralisApiKey KEY --moralisApiSecret SECRET --moralisSubdomain test.moralishost.com --autoSave 1 --moralisCloudfolder src\blockchain\cloud\

Some even up to 5 times. After we found this issue we cleared all the cloud functions and reran the script.wondering why this might have happened and how to prevent it in the future.

#2

what you mean by cloud function saved more than once?

#3

afterSave executed up to 4 times, where it should only have executed once. Issue has returned.

#4

you could set a flag so that the code doesn’t execute multiple times, afterSave should call when it is any update to that table rows

#5

thanks. looking into the request object, just strange that this is only happening now.

#6

is there any reason that a tables rows would be updating besides the handled confirm flag upon creation?

#7

important to note that this happens after an event is handled with eventsync

#8

do you have a nitro server now? (the one with coreservices)

in that case a create is also replaced with an update

#9

we have coreservices plugin, I believe that could be relevant to the behavior. however we are faced with a challenge, how can we accomplish managing a flag variable using event sync? Ideally we do not want to edit the table that receives the event. Also how does this factor into an afterSaving calling 3-4 times due to an event listener? The event synced is only producing 1 more row to the table.

#10

you could add a flag in the same table, it will trigger after save again

#11

we still do not know why the event is created/updated multiple times? If Moralis handles async functions in different threads, its possible that we will be unable to update the flag before a function is triggered again thus creating a race condition. Please advise if this scenario is a possibility

#12

you use an upgraded server?

you can also try to use redis if you want to implement some kind of mutex

#13

we use an upgraded server yes. rather than diving into a mutex implementation I am just trying to get a better understanding of how events are being synced (since the expectation is of a single thread to run rather than multiple threads in the first place)

#14

on what chain is the server? or on what chain are those events synced

also, what version of the server are you using, latest Nitro version?

#15

tested on 0.0.345 and moving to 0.0.347 to test again after facing some technical difficulties. How can I confirm that the server is Nitro? The Role table contains coreservices is how I currently check and have assumed that all updated servers are using Nitro.

#16

if you have coreservices plugin then it is a nitro server
only new created servers are nitro servers by default

#17

how do we confirm our current list of plugins/install coreservices plugin? Does not seem to be offered on the currently available plugin store and is not indicated in my installed plugins list (currently empty)

#18

it looks like that server that you have is not a nitro server

you can create a new server and test the event there, that new server will be a nitro server

1 Like
#19

hello, I have the same problem with a nitro server
my cloud function: Moralis.Cloud.beforeSave is called multiple times for the same event

#20

How many times it happens?