Logs that got sent sparsely or burstily would get buffered for long
periods of time since the logic to flush them only does so on the next
log write. In the worst case, a subsequent log write could never happen
and cause a log entry to be indefinitely buffered.
This fix introduces a recurring event/timer to simply flush all pending
logs at frequency of Broker::log_batch_interval.
During merge I split the test for bro_init/bro_done/bro_script_loaded
event errors into individual tests since the other testing of the zeek
versions of those events seemed fine to otherwise keep.
* origin/topic/robin/631-deprecation-v2:
Update NEWS for naming changes.
Small cleanup and updating submodules.
Remove test for legacy plugin.
Remove legancy symlinks in aux/.
Add warnings when loading scripts ending in ".bro", or using legacy environment variables.
Fix missing rename.
No longer symlink local.zeek to local.bro.
Update notice user agent.
Remove old_comm_usage_is_ok.
Remove bro-config.h.in and bro-path-dev.in.
Change Bro wrapper script to now abort when old executable names are still used.
Remove APIs that were explicitly deprecated to be removed in 3.1.
* origin/topic/jsiwek/supervisor: (44 commits)
Add note that Supervisor script APIs are unstable until 4.0
Move command-line arg parsing functions to Options.{h,cc}
Add btests for supervisor stem/leaf process revival
Move supervisor control events into SupervisorControl namespace
Fix supervisor "destroy" call on nodes not currently alive
Move supervisor source files into supervisor/
Address supervisor code re-factoring feedback from Robin
Convert supervisor internals to rapidjson
Add Supervisor documentation
Add supervisor btests
Improve logging of supervised node errors
Fix supervised node inheritence of command-line script paths
Improve normalize_path() util function
Use a timer to check for death of supervised node's parent
Improve supervisor checks for parent process termination
Improve handling of premature supervisor stem exit
Improve supervisor signal handler safety
Remove unused supervisor config options
Cleanup minor Supervisor TODOs
Improve supervisor debug logging
...
Typically in base scripts, Log::create_stream() is called in zeek_init()
handler with &priority=5 such that it will have already been created
in the default zeek_init() &priority=0.
Also have stem process execv() with original command-line arguments so
that they're re-parsed and inherited correctly by supervised-nodes in
the event the stem process needs to be re-created.
Adjustments during merge:
- kept the UNKNOWN Log::ID as placeholder value
- changed the coverage.find-bro-logs test to check for arbitrary $path
field values instead of just string literals
- don't force EnumVal to unsigned integer since the relevant union member
is the signed integer and added the relevant enum values/types to
.bif files for easier access
- compare FILE* versus file name to check for stdout equality (don't
think it matters much, just a bit more efficient)
- minor whitespace/style tweaks
* origin/topic/dev/print-to-log:
Added a non boolean configuration and other changes as suggested by Jon
Allow Print Statements to be redirected to a Log# This is a combination of 3 commits.
And switch Zeek's base scripts over to using it in place of
"connection_state_remove". The difference between the two is
that "connection_state_remove" is raised for all events while
"successful_connection_remove" excludes TCP connections that were never
established (just SYN packets). There can be performance benefits
to this change for some use-cases.
There's also a new event called ``connection_successful`` and a new
``connection`` record field named "successful" to help indicate this new
property of connections.
More aspects of the cluster configuration to get fleshed out later,
but a basic cluster like one would use for a live deployment
can now be instantiated and run under supervision. The new
clusterized-pcap-processing supervisor mode is also not done yet.
The process hierarchy and all supervisor control commands are now
working (e.g. status, create, destroy, restart), but nodes are
not currently spawned with the desired configuration parameters so
they don't yet operate as real cluster nodes (e.g. worker, logger,
manager, proxy).
* 'export_intel_events' of https://github.com/mauropalumbo75/zeek:
minor restyle and add comments
add an empty read_error event to the intel framework (in the export block, so that users can implement further checks with it)
move event Intel::read_entry to export block
Adjusted whitespace in merge.
* origin/topic/dev/non-ascii-logging:
Removed Policy Script for UTF-8 Logs
Commented out UTF-8 Script in Test All Policy
Minor Style Tweak
Use getNumBytesForUTF8 method to determine number of bytes
Added Jon's test cases as unit tests
Prioritizes escaping predefined Escape Sequences over Unescaping UTF-8 Sequences
Added additional check to confirm anything unescaping is a multibyte UTF-8 sequence, addressing the test case Jon brought up
Added optional script and redef bool to enable utf-8 in ASCII logs
Initial Commit, removed std::isprint check to escape
Made minor code format and logic adjustments during merge.
This allows one to tune the number of protocol violations to tolerate
from any given analyzer type before just disabling a given instance
of it.
Also removes the "disabled_aids" field from the DPD::Info record
since it serves no purpose: in this case, calling disable_analyzer
multiple times for the same analyzer is a no-op.
It may generally be better for our default use-case, as workers may
save a few percent cpu utilization as this policy does not have to
use any polling like the stealing policy does.
This also helps avoid a potential issue with the implementation of
spinlocks used in the work-stealing policy in current CAF versions,
where there's some conditions where lock contention causes a thread
to spin for long periods without relinquishing the cpu to others.
This signature is relevant for process dumps on Windows that could be extracted by various tools. The unencrypted transmission of the dump of a critical system process (for example, lsass.exe) via network would be detected by this rule.
In the past they were processed on the manager - which requires big
records to be sent around.
This has a potential of incompatibilities if someone relied on global
state for notice processing.
GH-214
