STAR Read Mapping on Platforma: cannot configure local storage path

Ask a question
1

I’m trying to configure Platforma to use a local storage path for data uploads, but I’m consistently encountering the following error:

rpc error: code = Unimplemented desc = Upload.Init: storage "main" is not accessible for data upload (driver "FS" does not support uploading)

I attempted to update the config.json file located at:

C:\Users\*user*\AppData\Roaming\platforma\config.json

I added the following “storage” block:

"storage": {
  "main": {
    "type": "host",
    "driver": "HostFS",
    "root": "C:\\Users\\*user*\\Platforma\\data"
  }
}

However, whenever I launch Platforma (from any method — desktop shortcut, direct exe, etc.), it overwrites the config.json file with default values, removing the custom “storage” section.

Setting the file to read-only doesn’t persist the configuration or enable uploads. As a result, any blocks that require file upload to storage: main fail with the above error.

Key Questions:
How can I permanently set or override the storage.main configuration to point to a local directory?

Is there a supported method for defining a custom config path or persistent storage driver?

Is HostFS expected to support uploading, or is another driver required for local usage?

2

Hi, if you run Platforma lockally you dont need to configure it, just select build in option:

image
image1920×1080 37.7 KB

If you run Platforma on a server, but your files are on a local machine you do not need to configure the storage locally, you can just drag and drop your files to Platforma and they will be copied to the server automatically.

3

Thank you for the response!

I am able to upload my raw .fastq files to Platforma when running it locally. The error occurs when I run the STAR Read Mapping block on my uploaded data.

I am able to run other blocks (e.g. MiXCR on TCR-seq data), so I think it may be specific to the STAR block.

Below is an example of the full error message for the featureCountsProgress process in the STAR block, but a similar error is generated for each process:

PlQuickJSError: PlErrorReport: resource: std/map:1, NG:0x93a/featureCountsProgress

PlInternalError: "NG:0x963" has 1 input errors:
[I] "NG:0x963/[\"NLFO62Z76XWAONOEDCP2UZSQ\"]": "NG:0x9B9" has 1 input errors:
[I] "NG:0x9B9/resource": "NG:0x9B8" has 1 input errors:
[I] "NG:0x9B8/resource": "NG:0x9B3" has 1 input errors:
[I] "NG:0x9B3/inputs": "NG:0x9B4" has 1 input errors:
[I] "NG:0x9B4/filesToAdd": "NG:0x9AF" has 1 input errors:
[I] "NG:0x9AF/sample_Aligned.sortedByCoord.out.bam": "NG:0x9AE" has 1 input errors:
[I] "NG:0x9AE/resource": "NG:0x9D5" has 1 input errors:
[I] "NG:0x9D5/resource": "NG:0x9D1" has 1 input errors:
[I] "NG:0x9D1/inputs": "NG:0x9D2" has 1 input errors:
[I] "NG:0x9D2/workdir": "NG:0x9CC" has 1 input errors:
[I] "NG:0x9CC/workdirIn": "NG:0x9C2" has 1 input errors:
[I] "NG:0x9C2/blobsIn": "NG:0x9C7" has 1 input errors:
[I] "NG:0x9C7/0x4000000000000002/blob": "NG:0x9E9" has 1 input errors:
[I] "NG:0x9E9/incarnation": rpc error: code = Unimplemented desc = Upload.Init: storage "main" is not accessible for data upload (driver "FS" does not support uploading)

QuickJS stacktrace:
PlErrorReport/uuid:4c8b3d04-4924-4457-b6ad-b808ee3d3266: 
    at resolveWithCommon (bundle.js)
    at resolve (bundle.js)
    at <anonymous> (bundle.js)
    at <anonymous> (bundle.js)
Host: PlErrorReport/uuid:4c8b3d04-4924-4457-b6ad-b808ee3d3266: 
    at QuickJSContext.unwrapResult (file:///C:/Users/user%20Lab/AppData/Local/Programs/platforma/resources/app.asar/node_modules/quickjs-emscripten-core/dist/chunk-JTKJZQYV.mjs:4:24443)
    at file:///C:/Users/user%20Lab/AppData/Local/Programs/platforma/resources/app.asar/node_modules/@milaboratories/pl-middle-layer/dist/index.mjs:2637:19
    at _Scope.withScope (file:///C:/Users/user%20Lab/AppData/Local/Programs/platforma/resources/app.asar/node_modules/quickjs-emscripten-core/dist/chunk-JTKJZQYV.mjs:4:402)
    at st.runCallback (file:///C:/Users/user%20Lab/AppData/Local/Programs/platforma/resources/app.asar/node_modules/@milaboratories/pl-middle-layer/dist/index.mjs:2632:97)
    at Object.___kernel___ (file:///C:/Users/user%20Lab/AppData/Local/Programs/platforma/resources/app.asar/node_modules/@milaboratories/pl-middle-layer/dist/index.mjs:3173:17)
    at L (file:///C:/Users/user%20Lab/AppData/Local/Programs/platforma/resources/app.asar/node_modules/@milaboratories/computable/dist/index.mjs:366:17)
    at E (file:///C:/Users/user%20Lab/AppData/Local/Programs/platforma/resources/app.asar/node_modules/@milaboratories/computable/dist/index.mjs:451:49)
    at I (file:///C:/Users/user%20Lab/AppData/Local/Programs/platforma/resources/app.asar/node_modules/@milaboratories/computable/dist/index.mjs:383:14)
    at E (file:///C:/Users/user%20Lab/AppData/Local/Programs/platforma/resources/app.asar/node_modules/@milaboratories/computable/dist/index.mjs:451:125)
    at I (file:///C:/Users/user%20Lab/AppData/Local/Programs/platforma/resources/app.asar/node_modules/@milaboratories/computable/dist/index.mjs:383:14)
4

Yes, the STAR block will not work on Windows when run locally unfortunately due to the way STAR software works. You can either run it locally on Mac/Linux or run Platforma on a server via ssh for example.