Arithmetic overflow panic in partition_tracker.rs:300 during scan operations (v2.0.0-alpha.4)

[2dev2]

Bug Description

Arithmetic overflow panic occurs during scan operations with partition filters.

Environment

• aerospike-core version: 2.0.0-alpha.4
• Rust version: [1.88]
• Operating System: macOS

Code to Reproduce

let pf = PartitionFilter::by_range(begin, count);
let policy = ScanPolicy::default();
let bins = Bins::from(&["dummy"][..]);

let recordset = aerospike_client
    .scan(&policy, pf, namespace, set_name, bins)
    .await?;

Expected Behavior

Scan should complete without panicking.

Actual Behavior

Thread panics with arithmetic overflow during partition tracking.


This affects the v2 async implementation. Please label as v2-related.

[khaf]

Thanks for your report. Can you share a little more information? How many records were there in the namespace/set? Do you have access to the stack trace of the crash?

[2dev2]

Thanks for looking into this.

1. We are using a single node cluster where we have 2Million records. There issue happens when system was running in long time and its very flaky. In scanning workflow we got this error.

2. Other time when we have like no records or very less records like 100 only. but our workflow we are checking global lock and if present or not based on that we were creating lock in that workflow we got above error.

I was not able to get stack trace, once we got stack trace we will share it.