parallelvirtualcluster
/
pvc
1
0
Fork 0
Code Issues 11 Pull Requests Projects Releases 111 Wiki Activity
2,338 Commits 1 Branch 113 Tags 28 MiB
Commit Graph

256 Commits

Author SHA1 Message Date
Joshua Boniface 284c581845 Ensure shutdown migrations actually time out 
Without this a VM that fails to respond to a shutdown will just spin
forever, blocking state changes.
 2021-06-15 00:23:15 -04:00
Joshua Boniface 953e46055a Fix issue with loading None version schema 2021-06-14 21:09:55 -04:00
Joshua Boniface d2bcfe5cf7 Bump version to 0.9.20 2021-06-14 18:06:27 -04:00
Joshua Boniface ef1701b4c8 Handle an additional exception case 2021-06-14 17:15:40 -04:00
Joshua Boniface 08dc756549 Actually disable the pvcapid service 
Prevents it from trying to start itself during updates or reboots on
non-primary coordinators.
 2021-06-14 17:13:22 -04:00
Joshua Boniface 0a9c0c1ccb Use a nicer reload method on hot schema update 
Instead of exiting and trusting systemd to restart us, instead leverage
the os.execv() call to reload the process in the current PID context.

Also improves the log messages so it's very clear what's going on.
 2021-06-14 17:10:21 -04:00
Joshua Boniface e34a7d4d2a Handle hot reloads properly 
A hot reload isn't possible due to DataWatch and ChildrenWatch
constructs, so we instead need to terminate the daemon to "apply" the
schema update. Thus we use exit code 150 (Application defined in LSB)
and reorder some of the elements of the schema validation to ensure
things happen in the right order.
 2021-06-14 12:52:43 -04:00
Joshua Boniface 1f49bfa1b2 Fix name of schema element 2021-06-13 20:56:17 -04:00
Joshua Boniface 647bce2a22 Ensure we don't grab None data 2021-06-13 16:43:25 -04:00
Joshua Boniface 26b1f531e9 Fix bad variable interpolation 2021-06-13 14:37:23 -04:00
Joshua Boniface be9f1e8636 Use more compatible is_alive in thread 2021-06-13 14:36:27 -04:00
Joshua Boniface b694945010 Fix incorrect name bug 2021-06-10 01:11:14 -04:00
Joshua Boniface 058c2ceef3 Convert VXNetworkInstance to new ZK schema handler 2021-06-10 00:36:18 -04:00
Joshua Boniface e7d60260a0 Fix typo in CephInstance path 2021-06-10 00:36:02 -04:00
Joshua Boniface 85aba7cc18 Convert VMInstance to new ZK schema handler 2021-06-09 23:15:08 -04:00
Joshua Boniface 7e42118e6f Adjust lock schema in NodeInstance and VMInstance 
Removes a superfluous lock and puts the sync_lock keys in more usable
places.
 2021-06-09 22:51:00 -04:00
Joshua Boniface 2704badfbe Convert VMConsole... to new ZK schema handler 2021-06-09 22:08:32 -04:00
Joshua Boniface 450bf6b153 Convert NodeInstance to new ZK schema handler 2021-06-09 22:07:32 -04:00
Joshua Boniface b94fe88405 Convert fencing to new ZK schema handler 2021-06-09 21:29:01 -04:00
Joshua Boniface 610f6e8f2c Convert CephInstance to new ZK schema handler 2021-06-09 21:17:09 -04:00
Joshua Boniface f913f42a6d Replace schema paths with updated zkhandler 2021-06-09 20:29:42 -04:00
Joshua Boniface e475552391 Fix some bugs with hot reload 2021-06-09 00:03:26 -04:00
Joshua Boniface 5540bdc86b Add automatic schema upgrade to nodes 
Performs an automatic schema upgrade when all nodes are updated to the
latest version.

Addresses #129
 2021-06-08 23:35:39 -04:00
Joshua Boniface 3c102b3769 Add per-node schema tracking 
This will allow nodes to start with their own schema versions, and then
be updated simultaneously by the API.

References #129
 2021-06-08 23:35:39 -04:00
Joshua Boniface a4aaf89681 Add ZKSchema loading and validation to Daemon 
Also removes some previous hack migrations from pre-0.9.19.

Addresses #129
 2021-06-08 23:35:39 -04:00
Joshua Boniface 5843d8aff4 Fix fence call to findTargetNode 2021-06-08 23:34:49 -04:00
Joshua Boniface cf96bb009f Bump version to 0.9.19 2021-06-06 01:47:41 -04:00
Joshua Boniface 719954b70b Fix missing list comma 2021-06-06 01:39:43 -04:00
Joshua Boniface 7dea5d2fac Move logger to common, fix buffering 2021-06-01 18:50:26 -04:00
Joshua Boniface 3a5226b893 Add missing flushed output 2021-06-01 18:30:18 -04:00
Joshua Boniface de2ff2e01b Fix removed function args 2021-06-01 17:02:36 -04:00
Joshua Boniface cd75413667 Increase initial lock timer 
With the new library the reader seems to be a little too quick, so hold
the write lock for 1 second instead of 1/2 second to ensure it is
caught.
 2021-06-01 17:00:11 -04:00
Joshua Boniface 9764090d6d Merge node common with daemon common 2021-06-01 12:22:11 -04:00
Joshua Boniface 12ac3686de Convert missed elements to new zkhandler 2021-06-01 11:57:21 -04:00
Joshua Boniface 5740d0f2d5 Remove obsolete zkhandler.py 2021-06-01 11:55:44 -04:00
Joshua Boniface 889f4cdf47 Convert common to new zkhandler 2021-06-01 11:55:32 -04:00
Joshua Boniface 8f66a8d00e Fix missed zkhandler conversion 2021-06-01 11:53:33 -04:00
Joshua Boniface 6beea0693c Convert fencing to new zkhandler 2021-06-01 11:53:21 -04:00
Joshua Boniface 1c9a7a6479 Convert VXNetworkInstance to new zkhandler 2021-06-01 11:49:39 -04:00
Joshua Boniface 790098f181 Convert VMInstance to new zkhandler 2021-06-01 11:46:27 -04:00
Joshua Boniface 8a4a41e092 Convert NodeInstance to new zkhandler 2021-06-01 11:27:35 -04:00
Joshua Boniface a48bf2d71e More gracefully handle none selectors 
Allow selection of "none" as the node selector, and handle this by
always using the cluster default instead of writing it in.
 2021-06-01 11:13:13 -04:00
Joshua Boniface a0b9087167 Set Daemon migration selector in zookeeper 2021-06-01 10:52:41 -04:00
Joshua Boniface 33a54cf7f2 Move configuration keys to /config tree 2021-06-01 10:48:55 -04:00
Joshua Boniface d6a8cf9780 Convert MetadataAPIInstance to new zkhandler 2021-05-31 19:55:09 -04:00
Joshua Boniface abd619a3c1 Convert DNSAggregatorInstance to new zkhandler 2021-05-31 19:55:01 -04:00
Joshua Boniface ef5fe78125 Convert CepnInstance to new zkhandler 2021-05-31 19:51:27 -04:00
Joshua Boniface f6d0e89568 Properly add absent node type 2021-05-31 19:26:27 -04:00
Joshua Boniface ede3e88cd7 Modify node daemon root to use updated zkhandler 2021-05-31 03:14:09 -04:00
Joshua Boniface 0c75a127b2 Bump version to 0.9.18 2021-05-23 17:23:10 -04:00
Joshua Boniface 9de14c46fb Bump version to 0.9.17 2021-05-19 17:06:29 -04:00
Joshua Boniface fe15bdb854 Bump version to 0.9.16 2021-05-10 01:13:21 -04:00
Joshua Boniface b851a6209c Catch all other exceptions in subprocess run 
Found a rare glitch where the subprocess pipes would not engage, causing
a daemon crash. Catch these exceptions with a retcode of 255 instead of
bailing out.

Closes #124
 2021-05-10 01:07:25 -04:00
Joshua Boniface 5ceb57e540 Handle emptying corrupted console log files 
Libvirt will someones write junk out to console log files, which breaks
the log parser deque with a UnicodeDecodeError.

If this happens, clear the log and re-open the deque again for newer
updates.

Closes #123
 2021-05-10 01:03:04 -04:00
Joshua Boniface 669338c22b Bump version to 0.9.15 2021-04-08 13:37:47 -04:00
Joshua Boniface c4ac75b973 Bump version to 0.9.14 2021-03-30 10:27:37 -04:00
Joshua Boniface 0bf276fd51 Update copyright year in headers 2021-03-25 17:01:55 -04:00
Joshua Boniface f4ec161aa2 Update file copyright header. 
Remove the option to select a later version of the GPL.
 2021-03-25 16:58:02 -04:00
Joshua Boniface 0ccfc41398 Bump version to 0.9.13 2021-02-17 11:37:59 -05:00
Joshua Boniface 9100c63e99 Add stored_bytes to pool stats information 2021-02-09 01:46:01 -05:00
Joshua Boniface aba567d6c9 Add nice startup banners to both daemons 
Add nicer easy-to-find (yay ASCII art) banners for the startup printouts
of both the node and API daemons. Also adds the safe loader to pvcnoded
to prevent hassle messages and a version string in the API daemon file.
 2021-02-08 02:51:43 -05:00
Joshua Boniface 0db8fd9da6 Bump version to 0.9.12 2021-01-28 16:29:58 -05:00
Joshua Boniface 9fbe35fd24 Bump version to 0.9.11 2021-01-05 15:58:26 -05:00
Joshua Boniface a24724d9f0 Use external ceph cmd for ceph df 2020-12-26 14:04:21 -05:00
Joshua Boniface 78c017d51d Remove erroneous extra colon in log output 2020-12-20 16:06:35 -05:00
Joshua Boniface 1b6613c280 Add live VNC information to domain output 
Sets in the node daemon, returns via the API, and shows in the CLI,
information about the live VNC listen address and port for VNC-enabled
VMs.

Closes #115
 2020-12-20 16:00:55 -05:00
Joshua Boniface d6ef722997 Fix bad log message 2020-12-15 10:51:52 -05:00
Joshua Boniface 518d699c15 Bump version to 0.9.10 2020-12-15 10:45:15 -05:00
Joshua Boniface ac3ef3d792 Revamp fencing order 
Prevents unnecessarily excessive timeouts if IPMI connections time out;
before, would have to go through 3 timed out commands at ~20s each
before failure was registered; reduced to 1 if the first times out.
 2020-12-15 02:48:25 -05:00
Joshua Boniface 3705daff43 Better handle failing RBD lock frees 
If the VM is not in a stop state, failing to free the lock is now
considered a fatal error and will put the domain into fail state,
aborting the start. This is better than being unsafe or trying to start
a VM which will fail to boot due to read-only volumes.
 2020-12-14 16:04:38 -05:00
Joshua Boniface 7c99a7bda7 Safely reset RBD locks on failed VMs 
Should correct issues on cold start as well as if a VM crashes
uncleanly, which would prevent the VM from starting due to stale RBD
locks.

This implementation has four parts:
  1. Update how IP addresses are handled, specifically by replacing all
  previous instances of "vni_ipaddr" with "vni_floatingipaddr", and then
  adding the "vni_ipaddr" with the real data for this node's IPs. Also
  include the storage IPs in this where they weren't before, so each
  this_node actually has the local IPs plus floating IPs. This enables
  the next two steps.
  2. Modify flush_locks to take this_node as an argument, and update the
  run_command function to only operate against this node, rather than on
  the primary coordinator.
  3. Have the flush_locks check each lock against the current node, to
  verify that the lock is actually held by the current node. This is the
  only way to do this safely. During fencing, we override this by not
  passing a this_node which bypasses this check.
  4. Have the VM start do the check for VM failure/startup and execute a
  flush_locks before actually starting the VM.
 2020-12-14 15:53:18 -05:00
Joshua Boniface 89c7e225a0 Move OSD stats uploading to primary only 
Instead of each node uploading its own OSD stats, which would not work
if the PVC daemon wasn't running, instead have the primary upload stats
for all OSDs in the cluster.
 2020-12-09 02:46:09 -05:00
Joshua Boniface b36ec43a2d Bump version to 0.9.9 2020-12-09 02:20:20 -05:00
Joshua Boniface ce5ee11841 Bump version to 0.9.8 2020-11-24 12:26:57 -05:00
Joshua Boniface d4a28d7a58 Bump version to 0.9.7 2020-11-19 10:48:28 -05:00
Joshua Boniface e69eb93cb3 Bump version to 0.9.6 2020-11-17 13:01:54 -05:00
Joshua Boniface 70dfcd434f Ensure inmigrate is cleared on failure 2020-11-17 12:57:37 -05:00
Joshua Boniface a4e5323e81 Bump version to 0.9.5 2020-11-17 12:34:04 -05:00
Joshua Boniface 9053edacd8 Bump version to 0.9.4 2020-11-10 15:33:50 -05:00
Joshua Boniface baac8f24fd Bump version to 0.9.3 2020-11-09 10:28:15 -05:00
Joshua Boniface 11702f4bc8 Bump version to 0.9.2 2020-11-08 02:03:29 -05:00
Joshua Boniface 6f66b77a00 Lint: E121/E126 continuation line under/over-indented for hanging indent 2020-11-07 15:06:21 -05:00
Joshua Boniface 9135c5e3e4 Lint: E241 multiple spaces after ',' 2020-11-07 14:52:39 -05:00
Joshua Boniface 260b39ebf2 Lint: E302 expected 2 blank lines, found X 2020-11-07 14:45:24 -05:00
Joshua Boniface ab0b932fe3 Lint: E125 continuation line with same indent as next logical line 2020-11-07 13:49:54 -05:00
Joshua Boniface f5988ad53d Lint: F821 undefined name 'pool'/'volume' 
This class is actually entirely unused but is kept for consistency with
the others. It may be used someday for something.
 2020-11-07 13:34:18 -05:00
Joshua Boniface c3dfe2e381 Lint: F821 undefined name 'myshorthostname' 2020-11-07 13:31:19 -05:00
Joshua Boniface 961ebb4c01 Lint: E305 expected 2 blank lines after class or function definition, found X 2020-11-07 13:17:49 -05:00
Joshua Boniface e553c5d42a Lint: E122 continuation line missing indentation or outdented 2020-11-07 13:12:26 -05:00
Joshua Boniface 7932be3948 Lint: E261 at least two spaces before inline comment 2020-11-07 13:11:03 -05:00
Joshua Boniface d2490419c5 Lint: E202 whitespace before ']' 2020-11-07 13:02:54 -05:00
Joshua Boniface d2e5ede399 Lint: E202 whitespace before ')' 2020-11-07 12:58:54 -05:00
Joshua Boniface 3f242cd437 Lint: E202 whitespace before '}' 2020-11-07 12:57:42 -05:00
Joshua Boniface b7daa8e1f6 E201 whitespace after '[' 2020-11-07 12:39:59 -05:00
Joshua Boniface c88965e898 Lint: E201 whitespace after '(' 2020-11-07 12:39:27 -05:00
Joshua Boniface e333f2b935 Lint: E201 whitespace after '{' 2020-11-07 12:38:31 -05:00
Joshua Boniface 3cb92fed75 Lint: E401 multiple imports on one line 2020-11-07 12:29:32 -05:00
Joshua Boniface 27c6ac2b66 Lint: W605 invalid escape sequence '\d' 
This is the only one where forcing an `r` type to the string was
required; the remainder of W605 were replaced with character class
enclosures.
 2020-11-07 12:22:20 -05:00
Joshua Boniface 8ba267a59e Lint: E211 whitespace before '['/'(' 2020-11-07 12:20:01 -05:00
Joshua Boniface 39cc992e9b Lint: E306 expected 1 blank line before a nested definition, found 0 2020-11-07 12:17:38 -05:00