Nothing shown after installing browser to local machine

[LongTianPy]

Hi,

I just installed the browser to my local machine following the instructions from http://wiki.wubrowse.org/Install_a_local_mirror, and I also put hg19.gz and hg19.gz.tbi into the seq folder. However, I got nothing when I clicked on the hg19 link on the index page. It says "Crashed loading genome info", and it would return to the index page with an error message saying "Unknown genome name: hg19". I'm not sure what is wrong with my installation step, but I was indeed confused at the part of enabling CORS: where should I add the three lines? Inside the block of the folder where the browser sits? Please help me with this issue, thank you!

[lidaof]

Hi @LongTianPy , CORS is an Apache setting, please check: https://poanchen.github.io/blog/2016/11/20/how-to-enable-cross-origin-resource-sharing-on-an-apache-server https://enable-cors.org/server_apache.html

do you load the makeDb.sql to hg19 too? Thanks.

[LongTianPy]

I think I've set CORS correctly because there has never been an error message about it.
And yes, I followed every step in the tutorial, including loading makeDb.sql to hg19, so that I have the config table in the hg19 database.

[LongTianPy]

Actually, as I'm installing this browser in a Microsoft Azure VM, there might be some dependencies other than those are listed in the tutorial that I didn't install. For example, does it require Perl?

[lidaof]

hi @LongTianPy , I don't think Perl is needed. can you do a git pull on our repo, on the config/hg19, replace the tracks.json and publichub.json on your server with the ones from the repo?

[LongTianPy]

I did git pull and replaced the json files, and it's still not working. When looking at the previous issues, I found that the apache error log of other people is

subtileKnife starting... >>>>>
cgi param parsed
Hooked to hg19
reviveScaffold() done

Whereas mine only shows

subtileKnife starting... >>>>>
cgi param parsed
Hooked to hg19

Does it mean that the database is successfully connected but the function of reviveScaffold is not run correctly?

[LongTianPy]

Although it says Hooked to hg19, the error returned on the screen says Unknown genome name: hg19. I wonder where these names are retrieved and connected.

[lidaof]

ok, weird, would you mind load the database again? under hg19 folder, do cat makeDb.sql | mysql -u hguser -p --local-infile=1 hg19

[LongTianPy]

I didn't have a problem loading the database but I did it once again just now. Still doesn't work.

In your sukn.js, line 308, onunknowngenome:on_unknown_genome, how is this parameter generated? I traced to line 466 of the same script, but I don't understand how it returns 'Unknown genome name'.

[lidaof]

Hi @LongTianPy , this is very strange,,actually I never encounter such issue before, I am not familiar with Microsoft Azure VM, are you using Ubuntu?

[LongTianPy]

Yes, I'm using Ubuntu 17.04 artful in the VM, only it was built from scratch, that's why I said that I might miss a couple of dependencies, e.g. Java? Also, is there any method I can check if the website is connected to hg19? When I click other tracks that I haven't implement, they showed the same error message ("Unknown genome name...") as I click hg19. So I guess there's something wrong with the database connection. Do you use any library to access MySQL with Javascript?

[lidaof]

the C code subtleKnife.c used mysql.h to bind mysql connection. Java is not needed. I am not sure if it's possible for you to generate a temporary user account for me to debug...as for now I cannot think of any particular reason...really sorry.

[LongTianPy]

I can create a user account but this VM can only be accessed within the internal network. But never mind I'll keep working on it. Thank you!

[lidaof]

We also have docker images in case you want to give it a try: https://hub.docker.com/r/epgg/eg/ Thanks.

[LongTianPy]

Thank you! I've worked it out. And I did try using the docker image, but MySQL failed to start.

[lidaof]

@LongTianPy ok, that;s great! Thanks.