1. 20 Oct, 2008 17 commits
  2. 18 Oct, 2008 2 commits
  3. 17 Oct, 2008 14 commits
  4. 16 Oct, 2008 7 commits
    • Iustin Pop's avatar
      Enable gnt-cluster modify to hv/beparams · 779c15bb
      Iustin Pop authored
      This patch enables the cluster modify to change:
        - enabled hypervisor list
        - hvparams (per hypervisor)
        - beparams (only the default group)
        gnt-cluster modify -B vcpus=3 -H xen-pvm:no_initrd_path
      Validation for parameters is somewhat missing - the individual
      hypervisors will be checked for syntax and validation, but beparams
      doesn't have validation yes (nowhere), it should be added here once we
      have a global method (will come soon).
      Reviewed-by: imsnah
    • Michael Hanselmann's avatar
      rapi: Convert to new HTTP server class · 16a8967d
      Michael Hanselmann authored
      Requests are no longer logged to a separate file.
      Reviewed-by: amishchenko
    • Guido Trotter's avatar
      design-2.0-cluster-parameters: escape final _ · c5e2d3b9
      Guido Trotter authored
      In rst a final _ is used for linking. Escaping it prevents an error at
      compile time on design-2.0-cluster-parameters.rst
      Reviewed-by: iustinp
    • Iustin Pop's avatar
      Show the cluster parametrs in gnt-cluster info · 469f88e1
      Iustin Pop authored
      This is just a raw update without any special formatting.
      Reviewed-by: imsnah
    • Oleksiy Mishchenko's avatar
      RAPI: Instance add/delete implementation. · 2f7635f4
      Oleksiy Mishchenko authored
      Reviewed-by: iustinp
    • Guido Trotter's avatar
      OS Design: invert new and old name · 29fc6636
      Guido Trotter authored
      Internally in ganeti we already have renamed the instance, so it's
      better to pass the new name in INSTANCE_NAME and the old one in
      OLD_INSTANCE_NAME rather than the old name in INSTANCE_NAME and the new
      one in NEW_INSTANCE_NAME as this reflects what the cluster thinks.
      Reviewed-by: iustinp
    • Iustin Pop's avatar
      Really fix with the zombie test (hopefully) · 09352fa4
      Iustin Pop authored
      So, instead of any timeouts for synchronization, we move to actual
      events. We still have one (huge) timeout for pathological cases (just to
      cleanup), but we don't sleep at all in normal usage.
      The patch moves the zombie/cleaned up child creation out from setUp(),
      and to on-demand by the functions. We remove all signal handler usage,
      instead we rely on the fact that when a process is cleaned up, its FDs
      are closed, so the read end of a pipe will get an event.
      Reviewed-by: imsnah