#24 Deploy task-libabigail
Closed: Fixed None Opened 7 years ago by mkrizek.

This will likely include:
adding whitelist of packages we trigger the check on to taskotron-trigger
create a cronjob that downloads this https://admin.fedoraproject.org/pkgdb/api/critpath which will be used as a whitelist
increase amount of ram for the minions
add the check repo to grokmirror to be pulled automatically (https://github.com/sinnykumari/task-libabigail)
* add the check into libtaskotron namespaces whitelist


I've increased the amount of ram and disk on the minons for dev to 6G and 40G using the testcloud config. If that goes well, we can move the config change to stg and prod

To get this in a format other than irc logs - my understanding of the initial whitelist is to use mostly critpath minus a few known issues for the size of clients we're running right now. Looking back at my logs, that list is:

  • webkit
  • llvm
  • thunderbird (and anything using xul)
  • java

I have a suspicion that the list will grow until we figure out how to get clients with more than 6G of memory but it's a start. @sinnykumari and @dodji - did I miss anything you know of on this list?

Also, is there a reason why the whitelist task was mashed together here instead of being broken off as its own task? Sure, the initial implementation is going to be pretty simple but this ticket is still a bit on the big-and-should-be-broken-up side.

! In #790#10870, @tflink wrote:
I've increased the amount of ram and disk on the minons for dev to 6G and 40G using the testcloud config. If that goes well, we can move the config change to stg and prod

Note that earlier today @jskladan increased the image to have 10G [1], I don't know how that goes together with the testcloud config, so just noting.

[1] https://infrastructure.fedoraproject.org/cgit/ansible.git/commit/?id=9c4b508722572ff287b7678fdced7230f9407960

Also, is there a reason why the whitelist task was mashed together here instead of being broken off as its own task? Sure, the initial implementation is going to be pretty simple but this ticket is still a bit on the big-and-should-be-broken-up side.

The reason being I was going to create the task for that but got distracted. :) I'll create it.

For the reference, noting the current configuration: 6GB RAM and 20GB (so minus the system, let's say 18GB) disk size.

Also, just a note, the task sends results into "world-writable" scratch namespace, we'll need to find a proper namespace for it when deploying to stg/prod.

Note that there is a issueswith triggering the task for each arch, it suffices to trigger just once, will investigate further.

Nevermind, it's triggered just once, but results are reported for each arch. Didn't notice it was the same job.

This has been deployed to dev, stg and prod. Closing, if there are any issues please open a new ticket.

Login to comment on this ticket.

Metadata