gather-users plugin reset itself and lost its list of members and previous meetings. about mattermost-gather-users

[juanfran]

which version are you using? If you are using v1 upgrade to v2 to save the data. The code to restore the data is here https://github.com/juanfran/mattermost-gather-users/blob/master/server/activate.go, what caused the reset? Maybe if you upgraded mattermost the info was lost.

[nough]

I'm using v2.1. We haven't made any changes to the Mattermost server. It has once again lost the activated users, I think as an overnight thing.

Our servers are updated and restarted each night, but we upgraded the mattermost server itself manually, and that hasn't been performed recently (it's definitely not the cause of the plugin data loss).

[nough]

Mattermost.log has about 500 of these errors, although sometimes it's "app/status.go:213" or 238
{"level":"warn","ts":1632986311.712406,"caller":"app/status.go:328","msg":"Failed to save status","user_id":"xxx","error":"failed in save Status: pq: duplicate key value violates unique constraint \"status_pkey\"","errorVerbose":"pq: duplicate key value violates unique constraint \"status_pkey\" failed in save Status github.com/mattermost/mattermost-server/v5/store/sqlstore.SqlStatusStore.SaveOrUpdate github.com/mattermost/mattermost-server/v5/store/sqlstore/status_store.go:50 github.com/mattermost/mattermost-server/v5/store/retrylayer.(*RetryLayerStatusStore).SaveOrUpdate github.com/mattermost/mattermost-server/v5/store/retrylayer/retrylayer.go:8303 github.com/mattermost/mattermost-server/v5/store/timerlayer.(*TimerLayerStatusStore).SaveOrUpdate github.com/mattermost/mattermost-server/v5/store/timerlayer/timerlayer.go:6894 github.com/mattermost/mattermost-server/v5/app.(*App).SaveAndBroadcastStatus github.com/mattermost/mattermost-server/v5/app/status.go:327 github.com/mattermost/mattermost-server/v5/app.(*App).SetStatusAwayIfNeeded github.com/mattermost/mattermost-server/v5/app/status.go:282 github.com/mattermost/mattermost-server/v5/app.(*WebConn).readPump.func2.1 github.com/mattermost/mattermost-server/v5/app/web_conn.go:281 github.com/mattermost/mattermost-server/v5/app.(*Server).Go.func1 github.com/mattermost/mattermost-server/v5/app/server.go:1133 runtime.goexit runtime/asm_amd64.s:1374"}

[nough]

as a further update: The server did not restart last night, and therefore kept its memory of who was assigned to meetings etc.

So I'm under the impression that data stored in our keyvaluestore using kvset and kvget etc are lost when the server is restarted. Unfortunately the mattermost api documentation doesn't state whether this is intended or not.

[nough]

Alejandro Garcia has recommended checking the contents of the table PluginKeyValueStore in the database to see if the data is still there but unreachable.

https://community.mattermost.com/core/pl/xzp7ezr9f3y5dbb8bgs5ubk9ew

[juanfran]

I really don't know what it can be, sorry.

One thing that can help is to restore the data manually. For example:

// get the user list
/gather-plugin info

// restore users
/gather-plugin add @user6 @user7 @user8

// previous meetings
/gather-plugin meetings

// restore previous meetings
/gather-plugin set_meetings {"user4":["user3"],"user1":[],"user2":["user4","user3"],"user3":["user4","user2"]}

[nough]

That would only work if I had saved copies of the meetings and users before the server restarted, but yes it is an option.

[nough]

Update:

• I don't believe this has happened in the last six months
• We have updated the mattermost server in the last six months, so it has definitely restarted (and will have restarted several times).

[nough]

I'm going to close this issue.