Commit Graph

356 Commits

Author SHA1 Message Date
Joshua Boniface 6ce28c43af Add MTU value checking and log messages
Ensures that if a specified MTU is more than the maximum it is set to
the maximum instead, and adds warning messages for both situations.
2021-10-09 18:48:56 -04:00
Joshua Boniface c45f8f5bd5 Have VXNetworkInstance set MTU if unset
Makes this explicit in Zookeeper if a network is unset, post-migration
(schema version 6).

Addresses #144
2021-10-09 17:52:57 -04:00
Joshua Boniface 3690a2c1e0 Fix migration bugs and invalid vx_mtu
Addresses #144
2021-10-09 17:35:10 -04:00
Joshua Boniface 50d8aa0586 Add handlers for client network MTUs
Refactors some of the code in VXNetworkInterface to handle MTUs in a
more streamlined fashion. Also fixes a bug whereby bridge client
networks were being explicitly given the cluster dev MTU which might not
be correct. Now adds support for this option explicitly in the configs,
and defaults to 1500 for safety (the standard Ethernet MTU).

Addresses #144
2021-10-09 17:02:27 -04:00
Joshua Boniface 6ee4c55071 Correct flawed conditional in verify_ipmi 2021-10-07 15:11:19 -04:00
Joshua Boniface c27359c4bf Bump version to 0.9.40 2021-10-07 14:42:04 -04:00
Joshua Boniface 46078932c3 Correct bad stop_keepalive_timer call 2021-10-07 14:41:12 -04:00
Joshua Boniface bdb9db8375 Bump version to 0.9.39 2021-10-07 11:52:38 -04:00
Joshua Boniface da9248cfa2 Bump version to 0.9.38 2021-10-03 22:32:41 -04:00
Joshua Boniface 23977b04fc Bump version to 0.9.37 2021-09-30 02:08:14 -04:00
Joshua Boniface f6f6f07488 Add timeouts to queue gets and adjust
Ensure that all keepalive timeouts are set (prevent the queue.get()
actions from blocking forever) and set the thread timeouts to line up as
well. Everything here is thus limited to keepalive_interval seconds
(default 5s) to keep it uniform.
2021-09-27 16:10:27 -04:00
Joshua Boniface 142c999ce8 Re-add success log output during migration 2021-09-27 11:50:55 -04:00
Joshua Boniface 1de069298c Fix missing character in log message 2021-09-27 00:49:43 -04:00
Joshua Boniface 55221b3d97 Simplify VM migration down to 3 steps
Remove two superfluous synchronization steps which are not needed here,
since the exclusive lock handles that situation anyways.

Still does not fix the weird flush->unflush lock timeout bug, but is
better worked-around now due to the cancelling of the other wait freeing
this up and continuing.
2021-09-27 00:03:20 -04:00
Joshua Boniface 0d72798814 Work around synchronization lock issues
Make the block on stage C only wait for 900 seconds (15 minutes) to
prevent indefinite blocking.

The issue comes if a VM is being received, and the current unflush is
cancelled for a flush. When this happens, this lock acquisition seems to
block for no obvious reason, and no other changes seem to affect it.
This is certainly some sort of locking bug within Kazoo but I can't
diagnose it as-is. Leave a TODO to look into this again in the future.
2021-09-26 23:26:21 -04:00
Joshua Boniface 3638efc77e Improve log messages during VM migration 2021-09-26 23:15:38 -04:00
Joshua Boniface c2c888d684 Use event to non-block wait and fix inf wait 2021-09-26 22:55:39 -04:00
Joshua Boniface febef2e406 Track status of VM state thread 2021-09-26 22:55:21 -04:00
Joshua Boniface 2a4f38e933 Simplify locking process for VM migration
Rather than using a cumbersome and overly complex ping-pong of read and
write locks, instead move to a much simpler process using exclusive
locks.

Describing the process in ASCII or narrative is cumbersome, but the
process ping-pongs via a set of exclusive locks and wait timers, so that
the two sides are able to synchronize via blocking the exclusive lock.
The end result is a much more streamlined migration (takes about half
the time all things considered) which should be less error-prone.
2021-09-26 22:08:07 -04:00
Joshua Boniface 3b805cdc34 Fix failure to connect to libvirt in keepalive
This should be caught and abort the thread rather than failing and
holding up keepalives.
2021-09-26 20:42:01 -04:00
Joshua Boniface 06f0f7ed91 Fix several bugs in fence handling
1. Output from ipmitool was not being stripped, and stray newlines were
throwing off the comparisons. Fixes this.

2. Several stages were lacking meaningful messages. Adds these in so the
output is more clear about what is going on.

3. Reduce the sleep time after a fence to just 1x the
keepalive_interval, rather than 2x, because this seemed like excessively
long even for slow IPMI interfaces, especially since we're checking the
power state now anyways.

4. Set the node daemon state to an explicit 'fenced' state after a
successful fence to indicate to users that the node was indeed fenced
successfully and not still 'dead'.
2021-09-26 20:07:30 -04:00
Joshua Boniface e23e2dd9bf Fix typo in log message 2021-09-26 03:35:30 -04:00
Joshua Boniface 0f02c5eaef Fix typo in sgdisk command options 2021-09-26 00:59:05 -04:00
Joshua Boniface 075abec5fe Use re.search instead of re.match
Required since we're not matching the start of the string.
2021-09-26 00:55:29 -04:00
Joshua Boniface 3a1cbf8d01 Raise basic exceptions in CephInstance
Avoids no exception to reraise errors on failures.
2021-09-26 00:50:10 -04:00
Joshua Boniface a438a4155a Fix OSD creation for partition paths and fix gdisk
The previous implementation did not work with /dev/nvme devices or any
/dev/disk/by-* devices due to some logical failures in the partition
naming scheme, so fix these, and be explicit about what is supported in
the PVC CLI command output.

The 'echo | gdisk' implementation of partition creation also did not
work due to limitations of subprocess.run; instead, use sgdisk which
allows these commands to be written out explicitly and is included in
the same package as gdisk.
2021-09-26 00:12:28 -04:00
Joshua Boniface 65df807b09 Add support for configurable OSD DB ratios
The default of 0.05 (5%) is likely ideal in the initial implementation,
but allow this to be set explicitly for maximum flexibility in
space-constrained or performance-critical use-cases.
2021-09-24 01:06:39 -04:00
Joshua Boniface d0f3e9e285 Bump version to 0.9.36 2021-09-23 14:01:38 -04:00
Joshua Boniface adc8a5a3bc Add separate OSD DB device support
Adds in three parts:

1. Create an API endpoint to create OSD DB volume groups on a device.
Passed through to the node via the same command pipeline as
creating/removing OSDs, and creates a volume group with a fixed name
(osd-db).

2. Adds API support for specifying whether or not to use this DB volume
group when creating a new OSD via the "ext_db" flag. Naming and sizing
is fixed for simplicity and based on Ceph recommendations (5% of OSD
size). The Zookeeper schema tracks the block device to use during
removal.

3. Adds CLI support for the new and modified API endpoints, as well as
displaying the block device and DB block device in the OSD list.

While I debated supporting adding a DB device to an existing OSD, in
practice this ended up being a very complex operation involving stopping
the OSD and setting some options, so this is not supported; this can be
specified during OSD creation only.

Closes #142
2021-09-23 13:59:49 -04:00
Joshua Boniface df277edf1c Move console watcher stop try up
Could cause an exception if d_domain is not defined yet.
2021-09-22 16:02:04 -04:00
Joshua Boniface 772807deb3 Bump version to 0.9.35 2021-09-13 02:20:46 -04:00
Joshua Boniface f3fb492633 Handle VM disk/network stats gathering exceptions 2021-09-12 19:41:07 -04:00
Joshua Boniface e962743e51 Add VM device hot attach/detach support
Adds a new API endpoint to support hot attach/detach of devices, and the
corresponding client-side logic to use this endpoint when doing VM
network/storage add/remove actions.

The live attach is now the default behaviour for these types of
additions and removals, and can be disabled if needed.

Closes #141
2021-09-12 19:33:00 -04:00
Joshua Boniface be954c1625 Don't crash cleanup if no this_node 2021-08-29 03:52:18 -04:00
Joshua Boniface fb46f5f9e9 Change default node object state to flushed 2021-08-29 03:34:08 -04:00
Joshua Boniface 694b8e85a0 Bump version to 0.9.34 2021-08-24 16:15:25 -04:00
Joshua Boniface a4c0e0befd Fix typo in output message 2021-08-23 00:39:19 -04:00
Joshua Boniface a18cef5f25 Bump version to 0.9.33 2021-08-21 03:28:48 -04:00
Joshua Boniface afb0359c20 Refactor pvcnoded to reduce Daemon.py size
This branch commit refactors the pvcnoded component to better adhere to
good programming practices. The previous Daemon.py was a massive file
which contained almost 2000 lines of direct, root-level code which was
directly imported. Not only was this poor practice, but this resulted
in a nigh-unmaintainable file which was hard even for me to understand.

This refactoring splits a large section of the code from Daemon.py into
separate small modules and functions in the `util/` directory. This will
hopefully make most of the functionality easy to find and modify without
having to dig through a single large file.

Further the existing subcomponents have been moved to the `objects/`
directory which clearly separates them.

Finally, the Daemon.py code has mostly been moved into a function,
`entrypoint()`, which is then called from the `pvcnoded.py` stub.

An additional item is that most format strings have been replaced by
f-strings to make use of the Python 3.6 features in Daemon.py and the
utility files.
2021-08-21 03:14:22 -04:00
Joshua Boniface afdf254297 Bump version to 0.9.32 2021-08-19 12:37:58 -04:00
Joshua Boniface 42e776fac1 Properly handle exceptions getting VM stats 2021-08-19 12:36:31 -04:00
Joshua Boniface 7ecc6a2635 Bump version to 0.9.31 2021-07-30 12:08:12 -04:00
Joshua Boniface 3ab6365a53 Adjust receive output to show proper source 2021-07-22 15:43:08 -04:00
Joshua Boniface 2a99a27feb Bump version to 0.9.30 2021-07-20 00:01:45 -04:00
Joshua Boniface fa1d93e933 Bump version to 0.9.29 2021-07-19 16:55:41 -04:00
Joshua Boniface 6ead21a308 Handle cleanup from a failure properly 2021-07-19 12:39:13 -04:00
Joshua Boniface b7c8c2ee3d Fix handling of this_node and d_domain in cleanup 2021-07-19 12:36:35 -04:00
Joshua Boniface d48f58930b Use harder exits and add cleanup termination 2021-07-19 12:27:16 -04:00
Joshua Boniface 7c36388c8f Add post-networking delay and adjust daemon delay 2021-07-19 12:23:45 -04:00
Joshua Boniface 71e4d0b32a Bump version to 0.9.28 2021-07-19 09:29:34 -04:00
Joshua Boniface 15d92c483f Bump version to 0.9.27 2021-07-19 00:03:40 -04:00
Joshua Boniface 602093029c Bump version to 0.9.26 2021-07-18 20:49:52 -04:00
Joshua Boniface b770e15a91 Fix final termination of logger
We need to do a bit more finagling with the logger on termination to
ensure that all messages are written and the queue drained before
actually terminating.
2021-07-18 19:53:00 -04:00
Joshua Boniface e23a65128a Remove del of logger item 2021-07-18 19:03:47 -04:00
Joshua Boniface 3a2478ee0c Cleanly terminate logger on cleanup 2021-07-18 18:57:44 -04:00
Joshua Boniface 323c7c41ae Implement node logging into Zookeeper
Adds the ability to send node daemon logs to Zookeeper to facilitate a
command like "pvc node log", similar to "pvc vm log". Each node stores
its logs in a separate tree under "/logs" which can then be combined or
queried. By default, set by config, only 2000 lines are kept.
2021-07-18 17:11:43 -04:00
Joshua Boniface cd1db3d587 Ensure node name is part of confing 2021-07-18 16:38:58 -04:00
Joshua Boniface 75fb60b1b4 Add VM list filtering by tag
Uses same method as state or node filtering, rather than altering how
the main LIMIT field works.
2021-07-14 00:59:20 -04:00
Joshua Boniface c6d552ae57 Rework success checks for IPMI fencing
Previously, if the node failed to restart, it was declared a "bad fence"
and no further action would be taken. However, there are some
situations, for instance critical hardware failures, where intelligent
systems will not attempt (or succeed at) starting up the node in such a
case, which would result in dead, known-offline nodes without recovery.

Tweak this behaviour somewhat. The main path of Reboot -> Check On ->
Success + fence-flush is retained, but some additional side-paths are
now defined:

1. We attempt to power "on" the chassis 1 second after the reboot, just
in case it is off and can be recovered. We then wait another 2 seconds
and check the power status (as we did before).

2. If the reboot succeeded, follow this series of choices:

    a. If the chassis is on, the fence succeeded.

    b. If the chassis is off, the fence "succeeded" as well.

    c. If the chassis is in some other state, the fence failed.

3. If the reboot failed, follow this series of choices:

    a. If the chassis is off, the fence itself failed, but we can treat
    it as "succeeded"" since the chassis is in a known-offline state.
    This is the most likely situation when there is a critical hardware
    failure, and the server's IPMI does not allow itself to start back
    up again.

    b. If the chassis is in any other state ("on" or unknown), the fence
    itself failed and we must treat this as a fence failure.

Overall, this should alleviate the aforementioned issue of a critical
failure rendering the node persistently "off" not triggering a
fence-flush and ensure fencing is more robust.
2021-07-13 17:54:41 -04:00
Joshua Boniface 2e9f6ac201 Bump version to 0.9.25 2021-07-11 23:19:09 -04:00
Joshua Boniface f09849bedf Don't overwrite shutdown state on termination
Just a minor quibble and not really impactful.
2021-07-11 23:18:14 -04:00
Joshua Boniface c76149141f Only log ZK connections when persistent
Prevents spam in the API logs.
2021-07-10 23:35:49 -04:00
Joshua Boniface f00c4d07f4 Add date output to keepalive
Helps track when there is a log follow in "-o cat" mode.
2021-07-10 23:24:59 -04:00
Joshua Boniface 20b66c10e1 Move two more commands to Rados library 2021-07-10 17:28:42 -04:00
Joshua Boniface cfeba50b17 Revert "Return to all command-based Ceph gathering"
This reverts commit 65d14ccd92.

This was actually a bad idea. For inexplicable reasons, running these
Ceph commands manually (not even via Python, but in a normal shell)
takes 7 * two orders of magnitude longer than running them with the
Rados module, so long in fact that some basic commands like "ceph
health" would sometimes take longer than the 1 second timeout to
complete. The Rados commands would however take about 1ms instead.

Despite the occasional issues when monitors drop out, the Rados module
is clearly far superior to the shell commands for any moderately-loaded
Ceph cluster. We can look into solving timeouts another way (perhaps
with Processes instead of Threads) at a later time.

Rados module "ceph health":
    b'{"checks":{},"status":"HEALTH_OK"}'
    0.001204 (s)
    b'{"checks":{},"status":"HEALTH_OK"}'
    0.001258 (s)
Command "ceph health":
    joshua@hv1.c.bonilan.net ~ $ time ceph health >/dev/null
    real    0m0.772s
    user    0m0.707s
    sys     0m0.046s
    joshua@hv1.c.bonilan.net ~ $ time ceph health >/dev/null
    real    0m0.796s
    user    0m0.728s
    sys     0m0.054s
2021-07-10 03:47:45 -04:00
Joshua Boniface 551bae2518 Bump version to 0.9.24 2021-07-09 15:58:36 -04:00
Joshua Boniface 2b5dc286ab Correct failure to get ceph_health data 2021-07-09 13:10:28 -04:00
Joshua Boniface 330cf14638 Remove return statements in keepalive collectors
These seem to bork the keepalive timer process, so just remove them and
let it continue to press on.
2021-07-09 13:04:17 -04:00
Joshua Boniface 65d14ccd92 Return to all command-based Ceph gathering
Using the Rados module was very problematic, specifically because it had
no sensible timeout parameters and thus would hang for many seconds.
This has poor implications since it blocks further keepalives.

Instead, remove the Rados usage entirely and go back completely to using
manual OS commands to gather this information. While this may cause PID
exhaustion more quickly it's worthwhile to avoid failure scenarios when
Ceph stats time out.

Closes #137
2021-07-06 11:30:45 -04:00
Joshua Boniface 7082982a33 Bump version to 0.9.23 2021-07-05 23:40:32 -04:00
Joshua Boniface 5b6ef71909 Ensure daemon mode is updated on startup
Fixes the side effect of the previous bug during deploys of 0.9.22.
2021-07-05 23:39:23 -04:00
Joshua Boniface be7b0be8ed Fix typo in schema path name 2021-07-05 23:23:23 -04:00
Joshua Boniface 37cd278bc2 Bump version to 0.9.22 2021-07-05 14:18:51 -04:00
Joshua Boniface a69105569f Add node PVC version data to Node information
Allows API client to see the currently-active version of the node
daemon.
2021-07-05 09:57:38 -04:00
Joshua Boniface 21a1a7da9e Fix bad schema reference
Not sure how this didn't cause an issue until now, but the wrong key
path was used and this was getting unexpected data with the newly-added
version string instead of the proper mode string.
2021-07-05 09:53:51 -04:00
Joshua Boniface f0fd3d3f0e Make extra sure VMs terminate when told
When doing a stop_vm or terminate_vm, check again after 0.2 seconds
and try re-terminating if it's still running. Covers cases where a VM
doesn't stop if given the 'stop' state.
2021-07-02 11:40:34 -04:00
Joshua Boniface f12de6727d Adjust logo slightly and add debug state 2021-07-02 02:32:08 -04:00
Joshua Boniface e94f5354e6 Update startup messages with new ASCII logo 2021-07-02 02:21:30 -04:00
Joshua Boniface c51023ba81 Add profiler to keepalive function 2021-07-02 01:55:15 -04:00
Joshua Boniface 39e82ee426 Cast base schema version to int
Or all our comparisons will fail later and nodes can't start.
2021-06-30 09:40:33 -04:00
Joshua Boniface fe0a1d582a Bump version to 0.9.21 2021-06-29 19:21:31 -04:00
Joshua Boniface 3490ecbb59 Remove explicit ZK address from Patronictl command 2021-06-22 03:31:06 -04:00
Joshua Boniface 2928d695c9 Ensure migration method is updated on state changes 2021-06-22 03:20:15 -04:00
Joshua Boniface 26dd24e3f5 Ensure MTU is set on VF when starting up 2021-06-22 02:26:14 -04:00
Joshua Boniface e623909a43 Store PHY MAC for VFs and restore after free 2021-06-22 00:56:47 -04:00
Joshua Boniface 60e1da09dd Don't try any shenannegans when updating NICs
Trying to do this on the VMInstance side had problems because we can't
differentiate the 3 types of migration there. So, just update this in
the API side and hope everything goes well.

This introduces an edge bug: if a VM is using a macvtap SR-IOV device,
and then tries to migrate, and the migrate is aborted, the NIC lists
will be inconsistent.

When I revamp the VMInstance in the future, I should be able to correct
this, but for now we'll have to live with that edgecase.
2021-06-22 00:00:50 -04:00
Joshua Boniface 7d42fba373 Ensure being in migrate doesn't abort shutdown 2021-06-21 23:28:53 -04:00
Joshua Boniface 24ce361a04 Ensure SR-IOV NIC states are updated on migration 2021-06-21 23:18:34 -04:00
Joshua Boniface eeb83da97d Add support for SR-IOV NICs to VMs 2021-06-21 23:18:22 -04:00
Joshua Boniface 64d1a37b3c Add PCIe device paths to SR-IOV VF information
This will be used when adding VM network interfaces of type hostdev.
2021-06-21 21:08:46 -04:00
Joshua Boniface 13cc0f986f Implement SR-IOV VF config set
Also fixes some random bugs, adds proper interface sorting, and assorted
tweaks.
2021-06-21 18:40:11 -04:00
Joshua Boniface ca11dbf491 Sort the list of VFs for easier parsing 2021-06-21 01:40:05 -04:00
Joshua Boniface e8bd1bf2c4 Ensure used/used_by are set on creation 2021-06-21 01:25:38 -04:00
Joshua Boniface bff6d71e18 Add logging to SRIOVVFInstance and fix bug 2021-06-17 02:02:41 -04:00
Joshua Boniface 57b041dc62 Ensure default for vLAN and QOS is 0 not empty 2021-06-17 01:54:37 -04:00
Joshua Boniface 5607a6bb62 Avoid overwriting VF data
Ensures that the configuration of a VF is not overwritten in Zookeeper
on a node restart. The SRIOVVFInstance handlers were modified to start
with None values, so that the DataWatch statements will always trigger
updates to the live system interfaces on daemon startup, thus ensuring
that the config stored in Zookeeper is applied to the system on startup
(mostly relevant after a cold boot or if the API changes them during a
daemon restart).
2021-06-17 01:45:22 -04:00
Joshua Boniface 8f1af2a642 Ignore hostdev interfaces in VM net stat gathering
Prevents errors if a SR-IOV hostdev interface is configured until this
is more defined.
2021-06-17 01:33:11 -04:00
Joshua Boniface e7b6a3eac1 Implement SR-IOV PF and VF instances
Adds support for the node daemon managing SR-IOV PF and VF instances.

PFs are added to Zookeeper automatically based on the config at startup
during network configuration, and are otherwise completely static. PFs
are automatically removed from Zookeeper, along with all coresponding
VFs, should the PF phy device be removed from the configuration.

VFs are configured based on the (autocreated) VFs of each PF device,
added to Zookeeper, and then a new class instance, SRIOVVFInstance, is
used to watch them for configuration changes. This will enable the
runtime management of VF settings by the API. The set of keys ensures
that both configuration and details of the NIC can be tracked.

Most keys are self-explanatory, especially for PFs and the basic keys
for VFs. The configuration tree is also self-explanatory, being based
entirely on the options available in the `ip link set {dev} vf` command.

Two additional keys are also present: `used` and `used_by`, which will
be able to track the (boolean) state of usage, as well as the VM that
uses a given VIF. Since the VM side implementation will support both
macvtap and direct "hostdev" assignments, this will ensure that this
state can be tracked on both the VF and the VM side.
2021-06-17 01:33:03 -04:00
Joshua Boniface 0ad6d55dff Add initial SR-IOV support to node daemon
Adds configuration values for enabled flag and SR-IOV devices to the
configuration and sets up the initial SR-IOV configuration on daemon
startup (inserting the module, configuring the VF count, etc.).
2021-06-15 22:56:09 -04:00
Joshua Boniface e4a65230a1 Just do the shutdown command itself 2021-06-15 02:32:14 -04:00