On Mon, Feb 05, 2018 at 09:27:15PM -0000, Andres Rodriguez wrote: > MAAS already has a mechanism to collapse retries into the initial request.
Are we certain that this is working correctly? If so, why are packet captures showing that MAAS is sending stacked tftp OACK responses, 1:1 for the duplicate incoming requests? It's clear to me that MAAS's handling at the wire level is incorrect - 10 retries of the same tftp request should result in a single OACK, not 10 of them (unless MAAS receives a retry *after* it has sent its OACK). I don't know if that also means that MAAS is inefficiently translating these into database requests on the backend. It had been suggested in this bug log and on IRC that MAAS *was* sending duplicate db requests for each of these packets; OTOH the timing of the stacked responses shows no latency in between them that would imply additional db round-trips. I think someone needs to directly inspect the behavior of a running MAAS server in this scenario to be sure. > In this case, it is the rack that grabs the requests and makes a request to > the region. If retries come within the time that the rack is waiting for a > response from the region, these request get "ignored" and the Rack will > only answer the first request. That is absolutely contradicted by the packet captures. The rack does not ignore the additional requests, it answers *ALL* of the requests. It's only the *client* that consolidates the duplicate responses from MAAS. (And then, because of a grub bug higher up the stack, re-requests the same file that it has already received.) -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1743249 Title: Failed Deployment after timeout trying to retrieve grub cfg To manage notifications about this bug go to: https://bugs.launchpad.net/maas/+bug/1743249/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
