Users don't automatically join auto-accept request communities when the owner of the community and requester go online

[VolodLytvynenko]

Description:

Users aren't automatically joining auto-accept request communities when both the community owner and the requester go online. The problem still occurs even without users going offline - online, although it happens less frequently in such cases. The steps below replicate this issue more consistently.

Steps:

1. User A is an owner of an auto-accept request community
2. User A shares community to User B
3. User A goes offline using Airplane mode
4. User B sends a request to join the community
5. User B goes offline
6. Wait 30 seconds - 1 minute
7. User A goes online
8. User B goes online
9. User B checks community status

Actual result:

User B is not joined to community. Community is still shown in the pending status

My.Movie.10.mp4

Expected result:

As soon as the community owner goes online all users requested to join the community automatically join the current community if current community is with auto-accept request

reference to feature related to this issue here

ENV:

• Nightly 18 Oct 2023

Additional info:

In my case, this issue is reproducible only when the community owner uses a real device (Pixel 7a, Android 13). But also it occurs during e2e tests running (here an example https://ethstatus.testrail.net/index.php?/tests/view/3261326)

Logs from physical devices:

• Community owner (User A) logs:
  Status-debug-logs.zip

• Community requester (User B) logs:
  logs.zip

Logs from e2e failed run:

• Community owner (Device 1) logs:
  logcat (1).log

• Community requester (Device 2) logs:
  logcat.log

[churik]

@VolodLytvynenko if User A goes offline on desktop, the issue is reproducible?

[VolodLytvynenko]

@VolodLytvynenko if User A goes offline on desktop, the issue is reproducible?

@churik The issue is no longer reproducible on mobile or desktop when 'User A' is offline. It seems that the last time it happened, there was a delivery issue. However, it's strange because it reproduced very easily earlier in every try even when message delivery worked perfectly for me.
I'll close this issue until re-investigate the steps just to be sure if the issue is no longer valid.

[VolodLytvynenko]

@VolodLytvynenko if User A goes offline on desktop, the issue is reproducible?

this issue is not reproducible if User A goes offline using a desktop in my case. Looks like this issue might be device-specific or happens randomly

[pavloburykh]

I have faced the issue with mobile admin node. Will check if the bug happens when admin node is Desktop.

[churik]

The same happens to me with admin on desktop (0.15 RC3 )

After admin goes back online, the request is Pending and the state is not changing until desktop re-logins

[qfrank]

FYI, i don't have such issue with local waku node, thought i can reproduce it with local waku node so i can go in deep 😅

[pavloburykh]

Comment from closed duplicate issue #18087 (comment)

As reported by @yevh-berdnyk

Problem

Steps:

• User A creates an open community and invites User B
• User A goes offline
• User B requests to join this community
• User A goes back online

Actual result:
Community is shown as pending endlessly

Expected result:
User B has joined the community
user_A_logcat.log.zip
user_B_logcat.log.zip
user_A_geth.log.zip
user_B_geth.log.zip
Note: As we can create communities only on e2e mobile builds I need also a desktop build with this status-go version to check this functionality

Related to PR : #17773

[pavloburykh]

Andrea's comment from closed duplicate issue #18087 (comment)

@siddarthkay

ERROR[12-05|04:05:07.134|github.com/status-im/status-go/protocol/messenger_mailserver_cycle.go:162]                                          Error getting new mailserver             error="error during polling loop: recvfrom: socket operation on non-socket"

[siddarthkay]

@VolodLytvynenko @pavloburykh can you check if this issue is still reproducible ?

[pavloburykh]

@VolodLytvynenko @pavloburykh can you check if this issue is still reproducible ?

@siddarthkay we have recently faced failed e2e because of this bug. So I believe it is still there. @VolodLytvynenko will provide more info on manual reproduction.

[siddarthkay]

Awesome! Thanks for confirming

[VolodLytvynenko]

@siddarthkay we have recently faced failed e2e because of this bug. So I believe it is still there. @VolodLytvynenko will provide more info on manual reproduction.

@pavloburykh @siddarthkay I can't reproduce this issue manually anymore. Previously, it was easy to do so, especially when users returned from offline, as described here: #17678 (comment)

[pavloburykh]

Thank you @VolodLytvynenko! @siddarthkay let's keep this issue open anyway. Once we have some update we will post it here.

[pavloburykh]

closing as issue is not reproducible anymore