Compressed extend table index doesn't match hash table size

[hddesai]

Hello,

After upgrading to dragmap version 1.3, I get the following error for one of my runs that previously had a segmentation fault:

The command line is as follows:
dragen-os -r dragmap_testref/ -1 mrna_S1_L001_R1_001.fastq.gz -2 mrna_S1_L001_R2_001.fastq.gz --output-directory mRNA_output/ --output-file-prefix mrna-alpha-x1

I am unsure what the exception means. Thank you for your help.

[mercywilliams160896]

Hi,
I am facing the same issue while I was trying to align my fastq files. Were you able to resolve the issue? If so, it would be really helpful if you could detail it. My error is as follows
(myenv) [mercy.rophina@node50 HGDP_Fastq]$ dragen-os -r /lustre/mercy.rophina/Ancestry_analysis/HGDP_Fastq/new_ref/ -1 ERR011001_R1_trimmed.fastq.gz -2 ERR011001_R2_trimmed.fastq.gz -v
2023-02-09 19:03:42 [2ab702654f80] Version: 1.3.0
2023-02-09 19:03:42 [2ab702654f80] argc: 8 argv: dragen-os -r /lustre/mercy.rophina/Ancestry_analysis/HGDP_Fastq/new_ref/ -1 ERR011001_R1_trimmed.fastq.gz -2 ERR011001_R2_trimmed.fastq.gz -v
decompHashTableCtxInit...
0.994 seconds
decompHashTableHeader...
0.002 seconds
decompHashTableLiterals...
2.191 seconds
decompHashTableExtIndex...
0.047 seconds
decompHashTableAutoHits...
Error: boost::exception: /opt/conda/conda-bld/dragmap_1659316645256/work/src/lib/reference/ReferenceDir.cpp(115): Throw in function dragenos::reference::ReferenceDir7::ReferenceDir7(const boost::filesystem::path&, bool, bool)
Dynamic exception type: boost::wrapexceptstd::logic_error
std::exception::what: Decompression populating automatic hit, target extend table slot nonempty

[hddesai]

Sorry I never figured it out

[IbrahimTanyalcin]

We have same error even after we recreate the hash tables using: dragen --build-hash-table.
The end result is always the same, after running dragen

Exception thrown in “path/to/reference_sender.cpp line 333” – Could not decompress path/to/hash_table.cmp – Compressed extend table index doesn’t match hash table size.

Do we at least have an idea where to start looking? The log files weren't that helpful.