Fix path to file in consul

[dstocco]

With the PR #1914 , the MID calibration run was changed in order to use QC files from apricot. However, this led to a bug since the path to consul had to be changed as well.
This PR should solve the issue, allowing to launch calibration runs at P2.

[github-actions]

REQUEST FOR PRODUCTION RELEASES:
To request your PR to be included in production software, please add the corresponding labels called "async-" to your PR. Add the labels directly (if you have the permissions) or add a comment of the form (note that labels are separated by a ",")

+async-label <label1>, <label2>, !<label3> ...

This will add <label1> and <label2> and removes <label3>.

The following labels are available
async-2023-pbpb-apass4
async-2023-pp-apass4
async-2024-pp-apass1
async-2022-pp-apass7
async-2024-pp-cpass0
async-2024-PbPb-apass1
async-2024-ppRef-apass1
async-2024-PbPb-apass2
async-2023-PbPb-apass5

[dstocco]

@shahor02 @chiarazampolli @davidrohr : can this be merged? The MID calibration is currently broken at P2 (sorry, my fault), so it would be good if this PR lands in P2 before the next flp-suite

[shahor02]

I am not familiar with the consul path conventions, in the O2DPG I see as many scripts setting QC_CONFIG="/o2/components/qc/ANY.... as QC_CONFIG="components/qc/ANY/....

@davidrohr @chiarazampolli, do you know the rule ?

[chiarazampolli]

I don't know but I assume that @dstocco is right, looking at the TPC scripts at least I see none with the /o2/ that Diego wants to remove. I also assume that he knows that this is the right way. I would approve.

[dstocco]

Well, on my side it is just try and fail. I've launched the calibration run by pointing to my local O2DPG in alihlt-gw-prod.cern.ch. If I use "/o2/" it fails. But if I launch it with the current PR, it works.
Unfortunately I did not find any documentation on this point...
In my previous PR I forgot that I could test it this way and simply assumed that I just needed to change the helper function.

[chiarazampolli]

Since Diego tested it, i merged it.

[chiarazampolli]

@knopers8 , @Barthelemy , can you confirm that like this is correct and point us to a documentation, if it exists, for the future?

[knopers8]

If the config file is retrieved from apricot by the QC executable, please refer to: https://github.com/AliceO2Group/QualityControl/blob/master/doc/Advanced.md#templating-config-files

The Apricot REST API, which appears to be directly used by O2DPG scripts, is documented at:

• https://github.com/AliceO2Group/Control/blob/master/apricot/docs/apricot_http_service.md
• <aliecs_hostname>:32188/docs/index.html

In particular, notice that when apricot is used directly, o2 has to be removed. It should be present in URLs passed to QC. I agree it's confusing.

[chiarazampolli]

Hi @knopers8 ,

Also to me it seems that the function uses the REST API (

O2DPG/DATA/common/gen_topo_helper_functions.sh

Line 280 in 691d4d5

add_QC_from_apricot()

)

From this page for the REST API (which I guess is what you refer to when you say "when apricot is used directly", right?):

https://github.com/AliceO2Group/Control/blob/master/apricot/docs/apricot_http_service.md

I'd deduce that the "o2" in front is not needed.

@dstocco , please confirm that everything is fine now, as soon as you can test it.

Chiara

[dstocco]

Thanks @knopers8 for the links. Indeed, in the previous PR I did not remove the /o2/, that is why it was failing.
@chiarazampolli : this time I managed in testing the modification at P2, and got no error. So, I'm pretty confident. Now the PR is merged, so, in principle, the fix should be included in the next FLP suite.

[knopers8]

From this page for the REST API (which I guess is what you refer to when you say "when apricot is used directly", right?):

Right.