Logs Show Written But Not Available In Queries

Using the .NET Realm library version 12.3.0
Running on Ubuntu Linux 24.04 64-bit Intel Xeon

I write a document to my local Realm instance. realm.Add(item true) returns the item that I wrote as expected.

I look in my Atlas sync logs.

  • I see the Sync-Write log has the _id of the document written…
  • There are 0 errors related to sync. Any errors in my logs are related to custom functions that are not tied to this sync.
"RealmReceipt": {
    "updated": [
      "ObjectID(\"66b27a341314b497d5a31e06\")",
      "ObjectID(\"66b24f9d1d5a4fcf4fb7f3a2\")"
    ]
  },

When I use Compass to query for the objectId, I get zero results.

Query - {_id : ObjectId('66b24f9d1d5a4fcf4fb7f3a2')}
Results - QUERY RESULTS: 0

Using Compass I dig into the various collections in the __realm_sync database.

I find the _id of the document in the state_RealmReceipt collection.

  • It does NOT match the state that I had written in my code. There are values in the state collection that should be different than what they are.
  • It does have the latestVersion field in the ____sync_meta_ collection that matches the version of the log entry shown in the history collection.
  • The documentId is NOT present in the unsynced_documents collection.

I have plenty of drive space on my server. It is running at 50% right now with lots of room. These write won’t impact drive space at all.

I cannot see ANY errors anywhere.

Help please.

Can you please send a link to your application in services.cloud.mongodb.com? I can take a look in the backend. As a quick heads up I will likely not get to it until tomorrow but will try to do then.

What is your email address? Or can I DM you here? Or do I have to post it to the forums?

You can DM it to me on the forums or post it here (only those with access AKA you or MongoDB employees with read-only permissions) will be able to view the page.

Alternatively, you can just send me the application_id which is the part of the URL:

https://services.cloud.mongodb.com/groups/${GROUP_ID}/apps/${APP_ID}

https://services.cloud.mongodb.com/groups/663eb8d3bdd74c40ab517724/apps/66a267cdaf9629d7ca455169/dashboard

Hi, most of the normal things I would look for (Compensating Writes, Permissions Errors, Corrective Erasures) don’t seem to be too present on the application (though there are a few compensating writes) and the changes seem to be flowing well across the system.

I did find one thing that is odd though and it stems from the fact that changes to and from Atlas are an asynchronous and separate process. So when you see the Sync->Write message, that means that the upload has been published to Sync History and from there it can be sent to other devices and is sent to a queue to be updated in MongoDB.

Generally this next step is on the order of milliseconds and is done asynchronously to ensure consistency with MongoDB. However, I do see a spike where the delay reached 10 hours before going back to 1 second for your application. We will look more into why that happened (and I have confirmed that oddly this was specific to just this one application and not a global issue).

In the meantime, I am curious:

  1. If the document is now in Atlas (from my metrics about 8 hours ago from when I am writing this is when thing got un-clogged)
  2. If this is something you are still observing or if you now see it more immediately in Atlas

Best,
Tyler

We are back up and running. Apparently there is a back-end Translator Service that went down and had to be restarted by the support team.

1 Like