1. 05 Jun, 2018 2 commits
  2. 04 Jun, 2018 4 commits
    • David Johnson's avatar
      Docker server-side core, esp new libimageops support for Docker images. · 66366489
      David Johnson authored
      The docker VM server-side goo is mostly identical to Xen, with slightly
      different handling for parent images.  We also support loading external
      Docker images (i.e. those without a real imageid in our DB; in that
      case, user has to set a specific stub image, and some extra per-vnode
      metadata (a URI that points to a Docker registry/image repo/tag);
      the Docker clientside handles the rest.
      
      Emulab Docker images map to a Emulab imageid:version pretty seamlessly.
      For instance, the Emulab `emulab-ops/docker-foo-bar:1` image would map
      to `<local-registry-URI>/emulab-ops/emulab-ops/docker-foo-bar:1`; the
      mapping is `<local-registry-URI>/pid/gid/imagename:version`.  Docker
      repository names are lowercase-only, so we handle that for the user; but
      I would prefer that users use lowercase Emulab imagenames for all Docker
      images; that will help us.  That is not enforced in the code; it will
      appear in the documentation, and we'll see.
      
      Full Docker imaging relies on several other libraries
      (https://gitlab.flux.utah.edu/emulab/pydockerauth,
      https://gitlab.flux.utah.edu/emulab/docker-registry-py).  Each
      Emulab-based cluster must currently run its own private registry to
      support image loading/capture (note however that if capture is
      unnecessary, users can use the external images path instead).  The
      pydockerauth library is a JWT token server that runs out of boss's
      Apache and implements authn/authz for the per-Emulab Docker registry
      (probably running on ops, but could be anywhere) that stores images and
      arbitrates upload/download access.  For instance, nodes in an experiment
      securely pull images using their pid/eid eventkey; and the pydockerauth
      emulab authz module knows what images the node is allowed to pull
      (i.e. sched_reloads, the current image the node is running, etc).  Real
      users can also pull images via user/pass, or bogus user/pass + Emulab
      SSL cert.  GENI credential-based authn/z was way too much work, sadly.
      There are other auth/z paths (i.e. for admins, temp tokens for secure
      operations) as well.
      
      As far as Docker image distribution in the federation, we use the same
      model as for regular ndz images.  Remote images are pulled in to the
      local cluster's Docker registry on-demand from their source cluster via
      admin token auth (note that all clusters in the federation have
      read-only access to the entire registries of any other cluster in the
      federation, so they can pull images).  Emulab imageid handling is the
      same as the existing ndz case.  For instance, image versions are lazily
      imported, on-demand; local version numbers may not match the remote
      image source cluster's version numbers.  This will potentially be a
      bigger problem in the Docker universe; Docker users expect to be able to
      reference any image version at any time anywhere.  But that is of course
      handleable with some ex post facto synchronization flag day, at least
      for the Docker images.
      
      The big new thing supporting native Docker image usage is the guts of a
      refactor of the utils/image* scripts into a new library, libimageops;
      this is necessary to support Docker images, which are stored in their
      own registry using their own custom protocols, so not amenable to our
      file-based storage.  Note: the utils/image* scripts currently call out
      to libimageops *only if* the image format is docker; all other images
      continue on the old paths in utils/image*, which all still remain
      intact, or minorly-changed to support libimageops.
      
      libimageops->New is the factory-style mechanism to get a libimageops
      that works for your image format or node type.  Once you have a
      libimageops instance, you can invoke normal image logical operations
      (CreateImage, ImageValidate, ImageRelease, et al).  I didn't do every
      single operation (for instance, I haven't yet dealt with image_import
      beyond essentially generalizing DownLoadImage by image format).
      Finally, each libimageops is stateless; another design would have been
      some statefulness for more complicated operations.   You will see that
      CreateImage, for instance, is written in a helper-subclass style that
      blurs some statefulness; however, it was the best match for the existing
      body of code.  We can revisit that later if the current argument-passing
      convention isn't loved.
      
      There are a couple outstanding issues.  Part of the security model here
      is that some utils/image* scripts are setuid, so direct libimageops
      library calls are not possible from a non-setuid context for some
      operations.  This is non-trivial to resolve, and might not be worthwhile
      to resolve any time soon.  Also, some of the scripts write meaningful,
      traditional content to stdout/stderr, and this creates a tension for
      direct library calls that is not entirely resolved yet.  Not hard, just
      only partly resolved.
      
      Note that tbsetup/libimageops_ndz.pm.in is still incomplete; it needs
      imagevalidate support.  Thus, I have not even featurized this yet; I
      will get to that as I have cycles.
      66366489
    • Leigh Stoller's avatar
      c5259a31
    • Leigh Stoller's avatar
      Fix a bug that was introduced when we shifted to using os_setup · e59fc714
      Leigh Stoller authored
      directly (on the Cloudlab clusters); we were losing a lock out that
      allowed DeleteSliver() to run while in the middle of a CreateSliver().
      This was resulting in a lot of email about node failures since the nodes
      were getting yanked out from underneath the CreateSliver(). From the
      user perspective, this did not matter much, since they wanted the slice
      gone, but it finally bothered me enough to look more closely.
      e59fc714
    • Leigh Stoller's avatar
      adefb6f5
  3. 31 May, 2018 2 commits
  4. 30 May, 2018 5 commits
    • Leigh Stoller's avatar
      Add support for linkwide properties which are far more efficient wrt the · aba79edd
      Leigh Stoller authored
      XML size on really big lans. I do not expect this to be used very often,
      but it is handy. On the geni-lib side:
      
      class setProperties(object):
          """Added to a Link or LAN object, this extension tells Emulab based
          clusters to set the symmetrical properties of the entire link/lan to
          the desired characteristics (bandwidth, latency, plr). This produces
          more efficient XML then setting a property on every source/destination
          pair, especially on a very large lan. Bandwidth is in Kbps, latency in
          milliseconds, plr a floating point number between 0 and 1. Use keyword
          based arguments, all arguments are optional:
      
              link.setProperties(bandwidth=100000, latency=10, plr=0.5)
      
          """
      aba79edd
    • Leigh Stoller's avatar
      Minor fixes to previous revision(s). · 428d54d3
      Leigh Stoller authored
      428d54d3
    • Leigh Stoller's avatar
      Several backend/RPC changes for reservations: · 8266ae51
      Leigh Stoller authored
      1. Return current set of reservations (if any) for a user when getting
         the max extension (piggy backing on the call to reduce overhead).
      
      2. Add RPC to get the reservation history for a user (all past
         reservations that were approved).
      
         Aside; the reservation_history table was not being updated properly,
         only expired reservations were saved, not deleted (but used)
         reservations, so we lost a lot of history. We could regen some of it
         from the history tables I added at the Portal for Dmitry, but not
         sure it is worth the trouble.
      
      3. And then the main content of this commit is that for both of the
         lists above, also return the experiment usage history for the project
         an dthe user who created the reservation. This takes the form of a
         time line of allocation changes so that we can graph node usage
         against the reservation bounds, to show graphically how well utilized
         the reservation is.
      8266ae51
    • Leigh Stoller's avatar
      Switch to graceful restart of apache instead of killing it outright, so · 735250d3
      Leigh Stoller authored
      that backend processes in flight do not get killed in their tracks. This
      might not work right, but elabinelab testing does not tell me much, lets
      see how it goes.
      735250d3
    • Leigh Stoller's avatar
  5. 23 May, 2018 1 commit
  6. 21 May, 2018 1 commit
  7. 15 May, 2018 1 commit
  8. 14 May, 2018 3 commits
  9. 07 May, 2018 3 commits
  10. 30 Apr, 2018 1 commit
  11. 26 Apr, 2018 1 commit
  12. 24 Apr, 2018 1 commit
  13. 23 Apr, 2018 1 commit
  14. 20 Apr, 2018 1 commit
    • Leigh Stoller's avatar
      Couple of small reservation system changes: · 636cea07
      Leigh Stoller authored
      1. Show the reason on the listing page, as popover in the last
         column. Makes it easier to approve directly from the listing page if
         you can see the reason.
      
      2. Add optional approval message to pass along to the user.
      636cea07
  15. 19 Apr, 2018 1 commit
    • Leigh Stoller's avatar
      Minor changes to update keys: · c9faef6e
      Leigh Stoller authored
      1. Do not mark down nodes as needing to be updated, they just get stuck
         there.
      
      2. In the CM daemon, what for slivers/aggregates that have been stuck in
         updating users for too long, and cancel the update. Typically this is
         cause a node is wedged or not participating (CORD), so again just
         cancel and reset the state.
      c9faef6e
  16. 18 Apr, 2018 2 commits
    • Leigh Stoller's avatar
      A tiny little tweak that allows you to set the IPs on layer 1 link · 728bd3bd
      Leigh Stoller authored
      ifaces that correspond to endpoints on nodes. Makes it easier to
      do something like this, if we init the interfaces on the nodes with IP
      and mask and bring it up.
      
      	# Add a raw PC to the request and give it an interface.
      	node1 = request.RawPC("node1")
      	iface1 = node1.addInterface()
      
      	# Specify the IPv4 address
      	iface1.addAddress(pg.IPv4Address("192.168.1.1", "255.255.255.0"))
      
      	# Add another raw PC to the request and give it an interface.
      	node2 = request.RawPC("node2")
      	iface2 = node2.addInterface()
      
      	# Specify the IPv4 address
      	iface2.addAddress(pg.IPv4Address("192.168.1.2", "255.255.255.0"))
      
      	# Add L1 link from node1 to node2
      	link1 = request.L1Link("link1")
      	link1.addInterface(iface1)
      	link1.addInterface(iface2)
      728bd3bd
    • Leigh Stoller's avatar
  17. 17 Apr, 2018 3 commits
  18. 16 Apr, 2018 1 commit
  19. 13 Apr, 2018 2 commits
  20. 09 Apr, 2018 2 commits
  21. 03 Apr, 2018 1 commit
    • Leigh Stoller's avatar
      When referring to a specific image version in the disk image URN, and · d83614dc
      Leigh Stoller authored
      that image is an *imported* system image, we are almost certainly
      referring to a version related to the origin image, not the local
      copy (which is not in sync with the origin wrt the history). In that
      case, we can use the hash from the image server to track down the local
      version.
      
      Aside; If I were going to redo the import mechanism, I would make image
      provenance the default for everyone, and make the option be whether to
      save or delete the image file (that was the entire reason I made
      provenance optional, most sites especially geni racks, do not have
      enough storage to maintain old images. And image import would always
      import the history too, which is something I did later when we thought
      deltas were going to solve our problems.
      d83614dc
  22. 29 Mar, 2018 1 commit
    • Leigh Stoller's avatar
      Reservations system changes: · df90d7a7
      Leigh Stoller authored
      1) Rework so that instead of relying on swapin__last + autoswap timeout,
         set expt_expires for classic experiments at the beginning of swapin
         time. This is cause swapin_last is not set till the end of swapin,
         and so during swapin the res system is in an inconsistent state since
         there is no way to determine when the experiment ends.
      
      2) On the Geni path, simplify expiration handling; do not allow a slice
         modification and expiration change at the same time; the bookkeeping
         and failure rollback is a pain, especially wrt reservation system,
         and this rarely ever actually happens, so get rid of a lot of
         complication.
      df90d7a7