Importing push tokens

Overview

The following steps are a recipe how to import the push tokens form csv file.

The needed of this script is requested in this ticket SUITEDEV-25068.

The recipe contains the commands while working on the issue

While the import script is working deletion worker duplicate-push-token-cleanup is working also for cleaning duplicate tokens imported

Description of the import-csv-v3

Link to the script import-csv-v3

This script reads a given CSV file and uses the me-client service API (V3 API) to register the client, assign the token to the client and connect the client to the anonymous contact.

The script can be used to import tokens for

mobile push (this is the default) and
web push (browsers).

Execution the script in the pod

kubectl run -it --image eu.gcr.io/ems-gap-images/me-push:latest me-push-<name-of-the-pod> -- bash

kubectl cp <import-csv-name>.csv me-push-<name-of-the-pod>:<import-csv-name>.csv

Inside the pod bash

node server/processes/scripts/import-csv-v3/index.js --filename <import-csv-name>.csv --bom true

The script will output imported rows imediately while execution

contactToken xxx
processRecord {
  platform: 'android',
  hardware_id: 'xxx-xxx',
  push_token: 'XXXX',
  application_code: 'XXXX-0000',
  language: 'en',
  last_app_activity: 'XX/XX/XXXX'
}
processRecord {
  platform: 'ios',
  push_token: 'XXXX',
  push_token: 'XXXX',
  application_code: 'XXXX-0000',
  language: 'en',
  last_app_activity: 'XX/XX/XXXX'
}

When the script is done, you will be informed with result like this

- Summary -
Time started: Wed Oct 28 2020 12:42:04 GMT+0000
Time finished: Wed Oct 28 2020 12:42:11 GMT+0000
Duration: 0 minutes
Valid Records in CSV: 47
Invalid Records in CSV: 0
Imported: 47
Skipped: 0
Errors: 0

Problems & Hints

Some of the records in the csv file might be without push_token. Those records will fail validation and be automatically filtered out.
Maybe some of the push tokens requests will failed while importing and you will get some of these 3 types of errors (ECONNREFUSED, ETIMEDOUT or Bad Gateway)

Steps for resolving the faced errors:

You need to find the csv file with errors inside the pod <import-csv-name>.csv-errors.csv

You need to generate new csv file with the failing records from the original file

for ETIMEDOUT or ECONNREFUSED

grep "ETIMEDOUT or ECONNREFUSED" <import-csv-name>.csv-errors.csv | cut -d ',' -f 4 > failed-ids.txt

fgrep -f failed-ids.txt <import-csv-name>.csv > failed-tokens.csv

for Bad Gateway

grep "Bad Gateway" <import-csv-name>.csv-errors.csv | cut -d ',' -f 3 > failed-ids.txt

fgrep -f failed-ids.txt <import-csv-name>.csv > failed-tokens.csv

We need to copy the first line of the original csv file to the new failed-tokens.csv
```
platform,hardware_id,push_token,application_code,language,last_app_activity
```
Run the script with the new file failed-tokens.csv