Details on a Failure of a client to sync to the server

Unless you are the type of person who love trivial details, I wouldn’t read this. This provides some detail about the failure of two clients to synchronize correctly to the server this morning. A record added yesterday did not sync to two different clients (one of which was my computer).

Problem:

A PanX user, Polly, opened database ‘Worker’. In it she did not find a record that a co-worker had added previously.

I opened the same database and the record did not appear on my computer. I did a a Download Data and the record then appeared on my computer.

From this, I conclude that there was a syncing error on Polly’s machine and on my machine. The record should have been added to both upon opening the database today.

Investigation:

(1) Addition of the record of interest:

The record was added to the DB yesterday, 2/18, at 9:14 am. It is server record ID 2047. I don’ have the timestamps from yesterday.

The record was edited at 4:56 pm.

Currently, (2/19 at 12:32 pm) the Timestamp is 763.

——-

(2). Addition of a later record.

Another record was added to the database at 4:56 pm. It is ID 2048, and currently has a timestamp of 752.

—————-

(3). Examination of Polly DB upon opening Worker this morning, when subject record was missing.

I obtained a copy of Polly DB before it was synced. I disconnected my computer from the server network, then opened Polly’s copy of the database.

Using this formula: aggregate({info(“serverrecordts”)},“max”), I obtained the maximum timestamp of 717.

(4). I did Download Data with my copy of Worker on my computer this morning and edited the subject record. The instrumentation log shows an edit around lines 2622-2666 of the instrumentation log, and record 2047 got a new timestamp of 756 (line 2667 of instrumentation log).

(4) I then connected Polly’s unsynced DB to the server network and connected the Worker DB.

Still no missing record.

Closed and re-opened Worker: still no missing record.

The Maximum Timestamp in the unsynced copy of the Worker DB is still 717, yet the server database has timestamps of 763 and higher.

———

(5) After doing a Download Data both records added yesterday have appeared on my computer.

Conclusion: The data confirms that the Worker DB did not sync to the server even though there were later changes. I don’t know why; I’ll leave that to Jim.