[BUG] Loading multilog fails after multiple restarts

[haal]

Describe the bug

Loading multilog after multiple restarts fails.

How to reproduce the bug?

The test case is:

1. start confluo using data path (e.g. --data-path=/var/db)
2. create multilog and write to it

RpcClient client = new RpcClient(HOST, PORT);
client.createAtomicMultilog(MULTILOG_NAME, "{ msg: STRING(8) }", StorageMode.DURABLE);
client.append("abcdefgh");
client.disconnect();

3. stop the confluo server
4. start confluo server using same data path as in step 1
5. check count of items - outputs 1

RpcClient client = new RpcClient(HOST, PORT);
client.loadAtomicMultilog(MULTILOG_NAME);
System.out.println(client.numRecords());
client.disconnect();

6. stop the confluo server
7. start confluo server using same data path as in step 1
8. retry step 5 - it fails with Floating point exception

Expected behavior

Step 8 should not fail and should output 1

Detailed logs

Error log: Floating point exception