Max records for a batch_write call is less than server's default

[nkonopinski]

The default server batch-max-requests size is 5k but this client's max appears to be 4096 set here:
https://github.com/aerospike/aerospike-client-python/blob/master/src/include/pool.h#L12

When I attempt to construct a batch larger than 4096 keys/records, I get 'Cannot allocate as_bytes'

records = br.BatchRecords(
    [br.Write(
        key=(myNamespace, mySet, myKey),
        ops=[op.write('', bytearray(value))],
     ) for key, value in items.items()
     ]
)
try:
    client.batch_write(records)
except (ex.AerospikeError, ex.ClientError, ex.ServerError) as e:
    log.error("Aerospike batch_write error: [{0}] {1}".format(e.code, e.msg))

Am I creating the batch incorrectly? Is there a way to get around the limitation without limiting batch_write() to 4096 records?

[juliannguyen4]

We are aware of this issue and have created a Jira ticket for it

[juliannguyen4]

This is currently being worked on here: #548

[anshulkamath]

Is there any update on the status of this PR?

[juliannguyen4]

Hi all, the PR is finished, and I will review it ASAP. Once the PR is merged, I can publish a pre-release build to PyPI to provide an early build with this change, but it will not be tested internally by our quality engineering team. Is this okay?

[anshulkamath]

A pre-release is fine for now! Is there a ballpark for how long it would take to complete internal testing? Thank you!

[juliannguyen4]

@anshulkamath Me and @DomPeliniAerospike are going to work on it today, but we don't have an ETA yet because there's a chance we might have to make major changes to the PR. We can update you again by tomorrow.

[anshulkamath]

Ah, makes sense. Thanks!

[juliannguyen4]

Our internal team has been working on this issue and it's still in-progress.