-
Notifications
You must be signed in to change notification settings - Fork 2
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Document mkngff workflow #2
Comments
The output sql shows the path.
It's hard-coded in the script at the moment but otherwise pretty arbitrary. The // in the output is also arbitrary but you are right we could print the target directory there. 👍🏼 |
On idr0125-pilot... Created Get Database password...
Use psql to get SECRET (last session ID):
SQL has paths like Create symlinks...
|
I don't know if that is a permissions issue? Trying those instructions...
Maybe that's also a permissions issue, but don't know what's needed there? |
Looking at the sql it has:
the
but I don't know what the |
That's the repositoryID. |
This is just because sudo is trying to change the directory and can't. If you have correct SQL, it wasn't fatal. |
This will mean it's the wrong secret key. |
With the instructions you provided for getting the secret key (at the top) it prints lots of info - including 3 session keys.
|
Part of the problem is that it's different on the pilots than it is from main servers. (It's based on the number of java processes that are started!) |
OK, so it's different on pilots from other servers. So what should I do? Trying this on
Followed same steps as above and got the same error...
So this seems to be not the right way to get the secret on the pilots. |
Sorry. The "limit 3" in the above is for a full server (with Indexer and PixelData process). Changing that to limit 1 should do the trick:
|
|
Getting Get link from https://merge-ci.openmicroscopy.org/jenkins/job/BIOFORMATS-build/499/default/artifact/bio-formats-build/ZarrReader/target/
But I still see ResourceError
|
The contents of your
but the error messages says:
i.e. there are two extra directory paths: "idr0054/zarr/" |
Looks better...
But I still see:
|
Hmmm..... did you possibly run the script twice? In the server log I see:
|
Trying with a different Fileset on idr0125-pilot...
Extrapolate the dirs we need for symlinks...
Run sql...
Rendering image...
|
Blitz log has same error as before:
|
Trying different server:
Need to mount BIA s3 on
Rendering fails - expected since we don't have latest ZarrReader...
Update ZarrReader on idr0138-pilot:
Hmm - seems that idr0138 doesn't have ZarrReader - can we just add it?
Try rendering again - now we get same error as on idr0125-pilot:
|
Testing execute
Checking image... http://localhost:1040/webclient/?show=image-3249534
4877818 ms is 81 minutes |
Test whether we need to add On idr0125-pilot, using the
http://localhost:1040/webclient/?show=well-1316985 ... This is viewable with File count: 21525 compared with
5296612 ms is 88 minutes. |
Testing https://github.com/yandex-cloud/geesefs as an alternative to goofys... On idr0125-pilot...
|
Installed Try running
|
Since 1pm on 6th Oct approx 67 hours -> 12 minutes per
|
NB This workflow has been moved to https://github.com/IDR/mkngff_upgrade_scripts (main README)
Once a submission has been processed by BioStudies, it will become available at a URL like:
https://uk1s3.embassy.ebi.ac.uk/bia-integrator-data/pages/S-BIAD815.html
We need to harvest the uuids from the links in the "Viewable Images" table. We can do this with the following JavaScript code, pasted into the
Console
tab of the browser dev tools:Which will print something like this:
That CSV is a table of
filesetName.ome.zarr, UUID
. We need to add the Fileset IDs from IDR to that table, usingidr-util
scripts from IDR/idr-utils#56That PR contains a file
idr_filesets.csv
which containsFileset ID, filesetName.ome.zarr
from IDR.It also contains a script to take the csv from above and add the appropriate Fileset IDs (from
idr_filesets.csv
).Checkout the
idr-utils
branch of that PR. This can be done on a local machine.Copy the csv generated by JavaScript above and save it into a file like
idr-utils/scripts/ngff_filesets/idr0051.csv
. You will see some examples included in that PR.Then run the script, passing in the IDR ID...
This will update the csv file you just created, adding in the Fileset IDs to a new 3rd column.
Now we want to use that data with
omero-mkngff
.We need to do everything as the
omero-server
user since we'll want to be able to create symlinks from the ManagedRep.E.g. working on
idr0138-pilot
...Created conda environment created as
omero-server
user, e.g.mkngff
and installed omero-py andomero-mkngff
Get Database password (and host) needed for psql, and set these to env variables. Also set variable for
$IDRID
so you can copy and paste other commands from below...Use psql to get SECRET (last session ID). NB: for pilot servers we only have 1 process (as in this example). For other servers, update the
1
to3
in this psql command:Copy the contents of
idr0051.csv
table from above (containsFileset ID
andUUID
) and create a copy of the csv in theomero-server
user's home dir...Now we can read that csv and create an sql file for each Fileset (named
FILESET_ID.sql
).In the loop below,
biapath
is likeS-BIAD815/51afff7c-eed4-44b4-95c7-1437d8807b97
anduuid
is like51afff7c-eed4-44b4-95c7-1437d8807b97
.The BIA s3 repository should be mounted under
/bia-integrator-data
:Check that e.g.
$ ls /bia-integrator-data/S-BIAD815/51afff7c-eed4-44b4-95c7-1437d8807b97/51afff7c-eed4-44b4-95c7-1437d8807b97.zarr
will give you0 OME
The
omero mkngff
command below also creates the symlinks we need, from the ManagedRepository to the s3-mounted data (if they don't already exist).Running sql on a different server (using saved sql)
Zip and copy sql to a different server.
Unzip and update the SECRET in all sql files, getting current
$SECRET
as aboveThe replace didn't work using
$SECRET
etc in the regex, so just use actual values...SECRETUUID
is the default placeholder if you didn't use--secret
option to create sql.We want to execute all sql, using the csv, and also to use
omero mkngff
to do just the symlink creation...Now we can try viewing the images in webclient.
NB: sometimes this can take a while for the memo file to be regenerated. To check on the timings you can use unique string from the fileset name
E.g.
2519114
ms is 42 minutes.The text was updated successfully, but these errors were encountered: