search

gvlproject / gvl.ansible.playbook

Playbook for building the Genomics Virtual Laboratory
7 stars 4 forks source link

db_key issue #91

Open igormakunin opened 7 years ago

igormakunin commented 7 years ago

I have a problem with db_key on GVL-4.2 deployed without any genome. I create a new db_key with data manager, but after Galaxy restart the new db_key is not visible to data managers. It is inconvenient, but probably is not a major issue that would prevent release of GVL-4.2

igormakunin commented 7 years ago

From what I see, data_manager_fetch_genome_dbkeys_all_fasta does not write to /mnt/galaxy/galaxy-app/tool-data/all_fasta.loc If I got it right, data managers look in /mnt/galaxy/galaxy-app/tool-data/all_fasta.loc for available genomes / db_keys. The issue was fixed by manual editing of this file.

jessicachung commented 7 years ago

I'm not able to recreate the issue. This is what I did:

Launched a 4.2 instance with blank indices.

Run the data manager "Create DBKey and Reference Genome" to fetch a new genome and create a dbkey. I used sacCer2.

This populated the loc files all_fasta.loc and dbkeys.loc in /mnt/galaxyIndices/tool-data/dm/toolshed.g2.bx.psu.edu/repos/devteam/data_manager_fetch_genome_dbkeys_all_fasta/16ab4dd6a0e5 successfully.

Checked the other data managers could see sacCer2 (they could).

Restarted Galaxy via Cloudman.

Checked data managers could run a sacCer2 job (bwa-mem indices were created successfully).

@igormakunin Could you list the steps you performed to recreate the error?

igormakunin commented 7 years ago

@jessicachung My steps: upload a fasta ref sequence into history, run "Create DBKey...", This populates /mnt/galaxy/galaxy-app/tool-data/dbkeys.loc Other data managers can see the new genome. Restart Galaxy. Data managers do not see the new genome. I fixed it by editing /mnt/galaxy/galaxy-app/tool-data/all_fasta.loc and /mnt/galaxyIndices/tool-data/dm/toolshed.g2.bx.psu.edu/repos/devteam/data_manager_fetch_genome_dbkeys_all_fasta/16ab4dd6a0e5/all_fasta.loc I used a big genome, 14 or 16 billion bases. XXLarge instance was deployed in QRISCloud using an old launcher.