I've got two problems that frequently come up and I could use some help.
1: Sometimes an agent on our build farm will act up and be the cause for a build failure. What a lot of developers are noticing (and correctly hating) is the fact that when TeamCity tries to run the build again, it does so on the same agent that previously failed. By default, all of our builds are clean builds, so there is no reason for TeamCity to have an affinity towards using the same build machine over and over for the same build, let alone a build that is coming across errors. Is there any way to convince TeamCity that, if it is trying to rebuild a previously failed task, it should do so on a different, compatible agent?
2: Recently Mercurial has been giving us a lot of grief. In 90% of the cases I've seen, TeamCity is unable to receive the proper information from Mercurial because uncompressed mode is checked in the VCS root's settings. 100% of the time, this is fixed by going into the VCS root and unchecking uncompressed transfer. Unfortunately, this is ALWAYS checked off by default. Is there any way to uncheck this in all of our VCS roots at once? Is there a way to ensure that it is not checked by default? Our developers are already starting to complain about the # of clicks it takes to set up a task for their build, having to add another won't be taken lightly.
Thanks for the help.