Open sanderdebacker opened 1 year ago
I have the exact same issue. I tried with reducing preload size with no luck. Anyone figured out? Thanks!
Command:
$ krakenuniq --db DBDIR --thread 1 --report-file REPORTFILE.tsv --preload-size 200M sample.fastq.gz
Output:
Warning: Overwriting REPORTFILE.tsv.
/home/biouser/miniconda3/envs/bioinfo/libexec/classify -d ./DBDIR/database.kdb -i ./DBDIR/database.idx -x 200M -r REPORTFILE.tsv -a ./DBDIR/taxDB -p 12
Database ./DBDIR/database.kdb
classify: unable to mmap ./DBDIR/database.kdb: Invalid argument
With the following command:
krakenuniq --db /mnt/f/KrakenUniq --threads 3 --output /mnt/f/SAMPLE.krakenuniq --preload-size 7G SAMPLE.fastq.gz
I get this output and error:
/home/visitor/bin/krakenuniq/classify -d /mnt/f/KrakenUniq/database.kdb -i /mnt/f/KrakenUniq/database.idx -t 3 -o /mnt/f/SAMPLE.krakenuniq -x 7G -a /mnt/f/KrakenUniq/taxDB -p 12
Database /mnt/f/KrakenUniq/database.kdb
classify: unable to mmap /mnt/f/KrakenUniq/database.kdb: Invalid argument
The directory KrakenUniq contains the database.kdb file as well as all the kmer_distrib files, database.kdb.counts, database.idx, and taxDB. These were donwloaded from https://benlangmead.github.io/aws-indexes/k2 as the main page suggests.
krakenuniq version = v1.0.2
Anyone encountered this before and knows why the database.kdb file is invalid?
Cheers!
The same error happened to me by using the standard krakenuniq database and --preload-size 500M
.
But the error disappears when using a 20 times smaller database.
So I suspect that --preload-size
didn't really play a role in solving the limited memory issue.
Hello - did you manage to solve this? I am currently having the same issue
Hello - did you manage to solve this? I am currently having the same issue
No, and since I haven't received any support here, I've moved on to other tools that have a more responsive community (Centrifuge for example).
This appears to have something to do with either your available RAM or your available disk space. I wish we had a support team, but we don't. @sanderdebacker, if you understood the challenges of supporting a tool with 1000s of users when the original developers (Derrick Wood, Florian Breitwieser) left my lab many years ago, and when NIH doesn't provide funds for ongoing support of software, perhaps you'd be a little less harsh.
With the following command:
krakenuniq --db /mnt/f/KrakenUniq --threads 3 --output /mnt/f/SAMPLE.krakenuniq --preload-size 7G SAMPLE.fastq.gz
I get this output and error:
/home/visitor/bin/krakenuniq/classify -d /mnt/f/KrakenUniq/database.kdb -i /mnt/f/KrakenUniq/database.idx -t 3 -o /mnt/f/SAMPLE.krakenuniq -x 7G -a /mnt/f/KrakenUniq/taxDB -p 12
Database /mnt/f/KrakenUniq/database.kdb
classify: unable to mmap /mnt/f/KrakenUniq/database.kdb: Invalid argument
The directory KrakenUniq contains the database.kdb file as well as all the kmer_distrib files, database.kdb.counts, database.idx, and taxDB. These were donwloaded from https://benlangmead.github.io/aws-indexes/k2 as the main page suggests.
krakenuniq version = v1.0.2
Anyone encountered this before and knows why the database.kdb file is invalid?
Cheers!