We lost the nice context-stack [jobname][taskname][...] at the beginning
of each log line when switching to logrus.
Define some field names that define these contexts.
Write a human-friendly formatter that presents these field names like
the solution we had before logrus.
Write some other formatters for logfmt and json output along the way.
Limit ourselves to stdout logging for now.
Implement
* pruning on source side
* local job
* test subcommand for doing a dry-run of a prune policy
* use a non-blocking callback from autosnap to trigger the depending
jobs -> avoids races, looks saner in the debug log
Done:
* implement autosnapper that asserts interval between snapshots
* implement pruner
* job pull: pulling + pruning
* job source: autosnapping + serving
TODO
* job source: pruning
* job local: everything
* fatal errors such as serve that cannot bind socket must be more
visible
* couldn't things that need a snapshotprefix just use a interface
Prefixer() instead? then we could have prefixsnapshotfilter and not
duplicate it every time...
* either go full context.Context or not at all...? just wait because
community climate around it isn't that great and we only need it for
cancellation? roll our own?
How it works:
`zrepl stdinserver CLIENT_IDENTITY`
* connects to the socket in $global.serve.stdinserver.sockdir/CLIENT_IDENTITY
* sends its stdin / stdout file descriptors to the `zrepl daemon` process (see cmsg(3))
* does nothing more
This enables a setup where `zrepl daemon` is not directly exposed to the
internet but instead all traffic is tunnelled through SSH.
The server with the source job has an authorized_keys file entry for the
public key used by the corresponding pull job
command="/mnt/zrepl stdinserver CLIENT_IDENTITY" ssh-ed25519 AAAAC3NzaC1E... zrepl@pullingserver
Don't use jobrun for daemon, just call JobDo() once, the job must
organize stuff itself.
Sacrifice all the oneshot commands, they will be reintroduced as
client-calls to the daemon.
The existing ByteStreamRPC requires writing RPC stub + server code
for each RPC endpoint. Does not scale well.
Goal: adding a new RPC call should
- not require writing an RPC stub / handler
- not require modifications to the RPC lib
The wire format is inspired by HTTP2, the API by net/rpc.
Frames are used for framing messages, i.e. a message is made of multiple
frames which are glued together using a frame-bridging reader / writer.
This roughly corresponds to HTTP2 streams, although we're happy with
just one stream at any time and the resulting non-need for flow control,
etc.
Frames are typed using a header. The two most important types are
'Header' and 'Data'.
The RPC protocol is built on top of this:
- Client sends a header => multiple frames of type 'header'
- Client sends request body => mulitiple frames of type 'data'
- Server reads a header => multiple frames of type 'header'
- Server reads request body => mulitiple frames of type 'data'
- Server sends response header => ...
- Server sends response body => ...
An RPC header is serialized JSON and always the same structure.
The body is of the type specified in the header.
The RPC server and client use some semi-fancy reflection tequniques to
automatically infer the data type of the request/response body based on
the method signature of the server handler; or the client parameters,
respectively.
This boils down to a special-case for io.Reader, which are just dumped
into a series of data frames as efficiently as possible.
All other types are (de)serialized using encoding/json.
The RPC layer and Frame Layer log some arbitrary messages that proved
useful during debugging. By default, they log to a non-logger, which
should not have a big impact on performance.
pprof analysis shows the implementation spends its CPU time
60% waiting for syscalls
30% in memmove
10% ...
On a Intel(R) Core(TM) i7-6600U CPU @ 2.60GHz CPU, Linux 4.12, the
implementation achieved ~3.6GiB/s.
Future optimization may include spice(2) / vmspice(2) on Linux, although
this doesn't fit so well with the heavy use of io.Reader / io.Writer
throughout the codebase.
The existing hackaround for local calls was re-implemented to fit the
new interface of PRCServer and RPCClient.
The 'R'PC method invocation is a bit slower because reflection is
involved inbetween, but otherwise performance should be no different.
The RPC code currently does not support multipart requests and thus does
not support the equivalent of a POST.
Thus, the switch to the new rpc code had the following fallout:
- Move request objects + constants from rpc package to main app code
- Sacrifice the hacky 'push = pull me' way of doing push
-> need to further extend RPC to support multipart requests or
something to implement this properly with additional interfaces
-> should be done after replication is abstracted better than separate
algorithms for doPull() and doPush()
config defines a single datastructure that can act both as a Map and as a Filter
(DatasetMapFilter)
Cleanup wildcard syntax along the way (also changes semantics).
Note the docs on the placeholder user property introduced with this
commit. The solution is not really satisfying but couldn't think of a
better one OTOMH
Job names are derived from job type + user-defined name in config file
CLI now has subcommands corresponding 1:1 to the config file sections:
push,pull,autosnap,prune
A subcommand always expects a job name, thus executes exactly one job.
Dict-style syntax also used for PullACL and Sink sections.
jobrun package is currently only used for autosnap, all others need to
be invoked repeatedly via external tool.
Plan is to re-integrate jobrun in an explicit daemon-mode (subcommand).
Pushing is achieved by inverting the roles on the established
connection, i.e. the client tells the server what data it should pull
from the client (PullMeRequest).
Role inversion is achieved by moving the server loop to the serverLoop
function of ByteStreamRPC, which can be called from both the Listen()
function (server-side) and the PullMeRequest() client-side function.
A donwside of this PullMe approach is that the replication policies
become part of the rpc, because the puller must follow the policy.