Error/Bug In running HaplotypeCaller

We’re running Parabricks after conversion to Singularity which is running alignment by fq2bam well.
I have an error that has cropped up in one sample while running haplotype caller, on a sample that was aligned with Parabircks fq2bam.

Here are the last few lines of the log file.
[PB Info 2024-Feb-20 09:05:43] chr5:18840001 41.7 5138485 123323
[PB Info 2024-Feb-20 09:05:53] chr5:22478401 41.8 5160897 123368
[PB Info 2024-Feb-20 09:06:03] chr5:26035201 42.0 5183307 123412
[PB Info 2024-Feb-20 09:06:13] chr5:29750401 42.2 5203777 123409
[PB ESC[31mErrorESC[0m 2024-Feb-20 09:06:15][src/genotype.cpp:223] allele out of index, expected idx < finalAlleles.size(), exiting.
[PB ESC[31mErrorESC[0m 2024-Feb-20 09:06:15][src/genotype.cpp:283] Unexpected negative qual value -0.000000, PLindex 0, expected qual >= 0, exiting.

The command being run was this.
pbrun haplotypecaller
–ref $my_reference
–in-bam sample_50.bam
–out-variants sample_50.haplo.g.vcf.gz
–haplotypecaller-options ‘-pcr-indel-model NONE --output-mode EMIT_VARIANTS_ONLY’
–gvcf --tmp-dir /scratch

My understanding is that by giving the .g.vcf.gz as part of the --out-variants option, the caller will create the compressed vcf file, and create an index file.
It seems like something is going wrong during the index creation. From the error, is the parabricks haplotypecaller making a call that can’t be indexed?

This hasn’t happened to other samples I’ve tested parabricks on so far.
Any help or advice would be greatly appreciated.