Commit graph

46 commits

Author SHA1 Message Date
Hannes Mehnert a08f35ee5e cleanups 2018-10-26 21:29:59 +02:00
Hannes Mehnert 811f3abc50 adjustments 2018-10-26 21:29:59 +02:00
Hannes Mehnert c399501a18 get rid of vm_config.vname 2018-10-26 21:29:59 +02:00
Hannes Mehnert 0441b8ab25 tls endpoint 2018-10-26 21:29:59 +02:00
Hannes Mehnert f939ff5a58 influx stats 2018-10-26 21:29:59 +02:00
Hannes Mehnert 1d4d7509dc remove vmm_wire, use asn.1 2018-10-26 21:29:59 +02:00
Hannes Mehnert 51a0344477 fix warnings 2018-10-26 21:29:59 +02:00
Hannes Mehnert 2239aafdb7 revive vmm_client 2018-10-26 21:29:59 +02:00
Hannes Mehnert bcb280aa00 refactor commands into vmm_commands 2018-10-26 21:29:59 +02:00
Hannes Mehnert efc043cd5c fix 'vmmc info' 2018-10-26 21:29:59 +02:00
Hannes Mehnert 182e2ae10c policies:
vmmc now has more subcommands
  - policy [-n name] returns all policies in name and below
  - add_policy [-n name] [--cpu cpuid] [--mem mem] [--bridge bridge] [--block size] adds a policy
  - remove [-n name] removes policy at name

policy is just the same which is in vmm_req_delegation, and vmm_resources now check them:
- you cannot insert a subpolicy violating the prefix
- you cannot insert a policy which would forbid current resource usage
- you cannot insert a policy with which any subpolicy would be invalid
- you can adjust (increase/decrease) a policy if the above invariants are kept

implement "force create" directly in vmmd: much nicer to
 - check resource constraints,
 - kill vm potentially,
 - and create a new vm,
all as single transaction.
2018-10-26 21:29:59 +02:00
Hannes Mehnert ea83013068 delegation -> policy 2018-10-26 21:29:59 +02:00
Hannes Mehnert e413b8c99a remove naming struggle in vm_config and Log.hdr 2018-10-26 21:29:59 +02:00
Stefan Grundmann 4c5a795a3b console fifos in separate directory 2018-10-07 01:24:33 +00:00
Stefan Grundmann 133884faf4 log, stats and console socket go in their own directory 2018-10-07 00:32:25 +00:00
Hannes Mehnert 38094a53e3 use vmm_trie in log and stat, cleanups 2018-09-28 22:44:38 +02:00
Hannes Mehnert 02f8d94db8 s/ukvm/hvt/ 2018-09-21 22:31:04 +02:00
Hannes Mehnert 38b98ab318 minor 2018-09-20 23:19:55 +02:00
Hannes Mehnert 99ba1c5e4b stats are back now! no longer two pullers, but now with one pusher :) 2018-09-20 22:53:42 +02:00
Hannes Mehnert e7b4742964 less is more, also unify default socket paths
and vmmc console command
2018-09-19 21:53:18 +02:00
Hannes Mehnert bd10209297 wip, vmmc and vmmd talk with each other! 2018-09-09 20:52:04 +02:00
Hannes Mehnert 9ec69e23cc rename Vmm_commands to Vmm_unix 2018-07-07 23:14:49 +02:00
Hannes Mehnert c9afb6f5dc CLOEXEC for all the openfile() calls to avoid leaking the fifo's (created and opened by vmmd itself) to ukvm-bin started later 2018-04-30 00:20:28 +02:00
Hannes Mehnert 2bf1b96178 get rid of FreeBSD kludge by using devfs! 2018-04-26 22:17:54 +02:00
Hannes Mehnert 0583fbfaf1 stats: transmit vmid in add/remove/stats, pid only in add
don't use /tmp anymore, but /var/run/albatross for fifos + sockets + vm images,
  and /var/db/albatross for ukvm-bin and crls, and /var/log/albatross for logging

vmm_console/vmm_log/vmm_stats_lwt: delete socket on startup if it exists

vmm_influxdb_stats: connects to vmm_stats socket and pushes every interval in
 influxdb line format via tcp to specified host and port
2018-04-26 00:03:46 +02:00
Hannes Mehnert 9696953cd7 revise force-restart: now with wait for kill and resource cleanup before start
allows to cleanup various hacks, such as checking for pid in vmm_resources
or removing temporarily the allocated resources from the resource map in vmm_engine

semantics is now slightly different, but for sure enhanced.
- each VM has a Lwt.wait () task attached in Vmm_engine.t (tasks : 'c String.Map.t)
- normal create shouldn't be much different, apart from memoizing the sleeper
- after waitpid is done in vmmd, and vmm_engine.shutdown succeeded, Lwt.wakeup is called for the sleeper
- force create now:
 - checks static policies
 - looks for existing VM (and task), if present: kill and wait for task in vmmd
 - continue with presence checking of vm name, dynamic policies, allocate resources (tap, img, fifo)

this means the whole randomness in filenames can be removed, and the
communication between vmm_console and vmm_client is working again (attach/detach
could not work since vmm_console knew only about "albatross.AAA.BBB.RANDOM",
whereas vmm_client insisted on "AAA.BBB"

resource overcommitment (and races in e.g. block device closing + opening) are
gone now, only if the old vm is cleanup up, resources for the new one are
allocated and it is executed
2018-04-05 01:02:45 +02:00
Hannes Mehnert a89b2925fd Vmm_core.cmd is now a variant (no longer polymorphic variant), some renames in Vmm_wire.Stats and Vmm_wire.Console to disambiguate 2018-04-04 22:16:31 +02:00
Hannes Mehnert fdab43aed6 vmm_engine: finish renaming Destroy_image -> Destroy_vm (broken since bb61388cfc) 2018-04-04 21:46:37 +02:00
Hannes Mehnert 7a4661b2e1 style: require lwt 3.0.0, fix warnings, disable 4 (fragile pattern matching) and 48 (implicit elimination of optional argument) 2018-04-03 22:58:31 +02:00
Hannes Mehnert a0c0f39734 vmm_stats: more debug, ignore vmmapi_open failure
vmm_stats_lwt: drop all pids on socket disconnect
vmmd: setup statistics slightly later (after the chmod on FreeBSD)
2018-04-01 23:59:12 +02:00
Hannes Mehnert c978dcfb14 less debug output of resources 2018-03-22 23:30:15 +01:00
Hannes Mehnert e48d3b3457 vmm_lwt/vmm_tls: less debug output of read/write stuff 2018-03-22 17:00:08 +01:00
Hannes Mehnert d3941e70c6 style 2018-03-22 17:00:08 +01:00
Hannes Mehnert b9d5fa94f9 call waitpid() earlier, separate freebsd kludge (chmod g+rw), fixes #9 2018-03-22 17:00:08 +01:00
Hannes Mehnert db8ae1ee37 compression, fixes #6 2018-03-22 17:00:08 +01:00
Hannes Mehnert cfa7ccd1e0 safer and clearer error semantics for all processes, fixes #5 2018-03-22 17:00:08 +01:00
Hannes Mehnert 88012094f8 remove unused Vmm_commands.waitpid, catch EINTR in call to Lwt_unix.waitpid (in Vmm_lwt) 2018-03-22 17:00:08 +01:00
Hannes Mehnert bb61388cfc new permission: force_create
a client certificate may either contain `Create or `Force_create permission.  If
the latter is used (vmm_req_vm --force), and a VM with the same name already
exists, this is destroyed (if the dynamic resources without the existing would
allow the new one to be deployed) and the new one is started.

I had this concrete deployment scenario, where kill ; create takes some minutes
since it is 10MB data which needs to be transferred from my laptop to a remote
server (me behind dialup).

- renamed `Image to `Create
- renamed `Destroy_image to `Destroy_vm
2018-03-22 17:00:08 +01:00
Hannes Mehnert 0052c3dc84 replace vmimage (Cstruct) with 0 after we wrote it to disk to avoid
the host process from referencing the whole vmimage
2018-03-22 13:54:33 +01:00
Hannes Mehnert 8807150c86 use a random temporary file instead of the serial numbers of the certificate chain for temporary host resources 2018-03-22 13:54:33 +01:00
Hannes Mehnert 8407d13b15 port to new asn + x509 2017-12-23 15:03:56 +01:00
Hannes Mehnert bd8cc0ad22 style 2017-12-20 22:29:22 +01:00
Hannes Mehnert 19c37d2ca1 pass --mem to ukvm-bin, now that is merged in master 2017-10-26 23:07:29 +02:00
Hannes Mehnert 1d3cf484c4 use SIGTERM, not SIGKILL - now solo5 ukvm has the ability to execute atexit functions 2017-10-26 22:21:15 +02:00
Hannes Mehnert bdedadf689 vmmd:
- fix fd leak (always close socket)
 - send first message (login) after renegotiation

vmm_stats:
 - remove unneeded functionality (keeping old statistics around)
 - translate internal tap names to bridge names
 - gather statistics from vmmapi as well

vmm_prometheus_stats:
 - new exporter of statistics to prometheus

*:
 - fix typo in README
 - style
2017-09-14 21:47:07 +01:00
Hannes Mehnert 02be3f4528 initial 2017-07-10 10:38:25 +01:00