MMseqs2: Clustering error due to invalid database read
Expected Behavior
After a plass assembly, the clustering of the various coassemblies.
mmseqs linclust -c 0.9 --min-seq-id 0.95 --cluster-mode 0 \
${output}/plass_assembly_noclust \
${output}/plass_ass_clust \
$scratch
Current Behavior
Raises an error stating the following:
Invalid database read for database data file=data/02_geneclust/plass_assembly_noclust, database index=data/02_geneclust/plass_assembly_noclust.index
Size of data: 132820104774
Requested offset: 6192668722089269
Invalid database read for database data file=data/02_geneclust/plass_assembly_noclust, database index=data/02_geneclust/plass_assembly_noclust.index
Size of data: 132820104774
Requested offset: 13511065749901201
Invalid database read for database data file=data/02_geneclust/plass_assembly_noclust, database index=data/02_geneclust/plass_assembly_noclust.index
Steps to Reproduce (for bugs)
Initial creation of the database with createdb:
mmseqs createdb data/plass_assembly/whole_assembly.faa \
${output}/plass_assembly_noclust
MMseqs Output (for bugs)
Error raised: https://gist.github.com/adriaaula/92582186d55e5a8575420c6e8f8ce7e1
General output: https://gist.github.com/adriaaula/ed7062dffd5f7ca2f498bed2fa3217db
Context
Clustering of a big fasta with 4 independent coassemblies of ~8 samples each. 250 Gb of RAM and 48 cpus.
Your Environment
Include as many relevant details about the environment you experienced the bug in.
- Git commit used: d36dea
- Which MMseqs version was used: statically compiled (Statically-compiled, self-compiled, Homebrew, etc.): statically compiled
- Server specifications (especially CPU support for AVX2/SSE and amount of system memory):AVX2 supported
- Operating system and version:
About this issue
- Original URL
- State: closed
- Created 7 years ago
- Comments: 17 (9 by maintainers)
Commits related to this issue
- Rework memory allocation in case of split for kmermatcher https://github.com/soedinglab/MMseqs2/issues/208 — committed to soedinglab/MMseqs2 by martin-steinegger 7 years ago
Thanks, I don’t need the output of the awk call anymore, since I tracked down whats going wrong. It just a bit tricky to fix correctly and not waste a lot of RAM in doing so.