UPGRADE

UPGRAGE guidelines.
pmacct is developed keeping an eye to backward compatibility: the upgrade to
some newer version should be as smooth as possible from an user standpoint.
However, sometimes the upgrade may require some operations aimed to support
the changes done or break old assumptions no longer valid; while the effort
is to keep these cases at a low, please read this file in preparation to
upgrading your installation.

TO:	>= 1.7.9
FROM:	<= 1.7.8
TOPIC:	NetFlow v9 / IPFIX and flows with (ingress / egress) VRF ID field
DESC:   Version 1.7.9 does formalize much more than previous releases the
	presence of an ingress / egress VRF ID field as part of a NetFlow
	v9 / IPFIX packet: if such field is present, the corresponding RD for
	the VRF is looked up in the BGP feed for correlation, failing which,
	any BGP primitive in the aggregation method will remain unpopulated.

	Symptoms, examples and recommended remedies are documented in two (at
	the time of writing) issues on GitHub:
	* https://github.com/pmacct/pmacct/issues/768
	* https://github.com/pmacct/pmacct/issues/777

	In the case where a platform reports an ingress / egress VRF ID but a
	lookup in the Global Routing Table is required, a flow_to_rd_map can
	be configured. For example a line "id=0:0:0 ip=10.0.0.1" would look
	up flows coming from network element 10.0.0.1 in the GRT. For further
	guidance look examples/flow_to_rd.map.example.

TO:	>= 1.7.9
FROM:	<= 1.7.8
TOPIC:	nfacctd_disable_checks, sfacctd_disable_checks
DESC:   Due to newly introduced nfacctd_pre_processing_checks config knob,
	it was felt that nfacctd_disable_checks needed to be renamed for a
	more clear disambiguation. While the old name will still be available
	and will only be deprecated over time, it has been removed from docs 
	and renamed nfacctd_disable_sanity_checks.

TO:	>= 1.7.9
FROM:	<= 1.7.8
TOPIC:	imt_passwd
DESC:	This configuration knob was removed from user-facing documentation
	long ago but underlying code was never deprecated. With this release
	the feature is effectively not supported anymore.

TO:	>= 1.7.9
FROM:	1.7.8
TOPIC:	vlan, in_vlan primitives
DESC:	Due to a bug, in release 1.7.8 only, using the 'vlan' primitive did
	result in being using the 'vlan_in' field on output. Issues were
	opened on GitHub, ie. https://github.com/pmacct/pmacct/issues/696 .
	The old behaviour, pre-1.7.8 was restored, where 'vlan' primitive
	does map to 'vlan' on output; in a straightforward fashion 'in_vlan'
	primitive does now map to 'vlan_in' output field and 'out_vlan' to
	'vlan_out'. 

TO:	>= 1.7.8
FROM:	<= 1.7.7
TOPIC:	kafka_avro_schema_registry
DESC:	Until version 1.7.7 a custom "-value" string was forcingly appended 
	to the schema name. This was non-standard practice and hence such
	practice has been terminated. 

TO:     >= 1.7.8
FROM:   <= 1.7.7
TOPIC:	mpls_stack_depth primitive
DESC:	Support for the primitive has been removed. This has been coincidental
	with the introduction of the mpls_label_stack primitive, that exposes
	the full MPLS label stack, including its complex representation (on
	top of its stringy one), via the mpls_label_stack_encode_as_array
	configuration directive, in JSON and Apache Avro formats.

TO:	>= 1.7.8
FROM:	<= 1.7.7
TOPIC:	BGP, BMP message log / dumps & "label" field
DESC:	Until version 1.7.7, the output of BGP message logs / dumps would
	include MPLS labels, ie. as a result of a NLRI being part of a VPN,
	in the "label" field for both JSON and Avro encodings. In order to
	align naming convention with flow data plugins and to facilitate
	introduction of pre_tag_map support in BGP / BMP daemons (which
	output can be a "tag" or a "label" field), MPLS labels will now be
	encoded in the "mpls_label" field.  

TO:	>= 1.7.7
FROM:	<= 1.7.6
TOPIC:	pmtelemetryd, JSON telemetry input and JSON daemon output
DESC:	Until version 1.7.6, JSON telemetry input data would be embedded as
	a 'telemetry_data' string part of the JSON daemon output along with
	some other meta-data. While this was useful, it was slightly not
	convenient as extracting actual content from 'telemetry_data' does
	require some (simple) scriptware in the middle; starting version
	1.7.7, 'telemetry_data' is a JSON object (hence making the whole
	JSON daemon output nested) making it easier for consumers to parse
	the content.

TO:	>= 1.7.7
FROM:	<= 1.7.6
TOPIC:  Obsoleted features
DESC:   Following is the list of features, knobs and plugins that are being
        discontinued with release 1.7.7:
	* sql_history_since_epoch was removed from docs for a while; it is
	  replaced by timestamps_since_epoch so that all timestamps-related
	  config keys have the 'timestamps' suffix;
	* refresh_maps and pre_tag_map_entries were removed from docs for a
	  while; they are replaced respectively by maps_refresh and
	  maps_entries so that all maps-related config keys have the 'maps'
	  suffix.
	* interface_wait and interface were removed from docs for a while;
	  they are replaced respectively by pcap_interface_wait and
	  pcap_interface so that all pmacctd / libpcap-related config keys
	  have the 'pcap' suffix.

TO:	>= 1.7.6
FROM:	<= 1.7.5
TOPIC:	Obsoleted features
DESC:   Following is the list of features, knobs and plugins that are being
        discontinued with release 1.7.6:
	* Prefix labels, enabled with --enable-plabel configure switch, has
	  been removed as now pre_tag_map supports a 'set_label' action to
	  achieve the same.
	* sfacctd & timestamp_start aggregation primitive: sFlow does not
	  include a timestamp as part of its header structure and a time is
	  effectively created on arrival (timestamp_arrival). While this is
	  still the case and timestamp_arrival is populated, timestamp_start
	  will cease to be aliased over timestamp_arrival in order to avoid
	  confusions.

TO:     >= 1.7.6
FROM:   <= 1.7.5
TOPIC:  Default nfprobe plugin protocol version
DESC:   Since its conception nfprobe plugin default export version has been
	5. With IPFIX having been standardized back in 2013 with RFC7011 and
	being widely adopted by both exporters and collector, it is felt it
	is the right moment to switch the default version. On a sidenote,
	there are no plans to retire support for verdion 5.

TO:	>= 1.7.6
FROM:	<= 1.7.5
TOPIC:	BMP dump timestamps
DESC:	In order to align the namespace of timestamps to existing ones, ie.
        timestamp_start timestamp_end timestamp_arrival etc., the field
	'event_timestamp' was renamed 'timestamp_event'.

TO:	>= 1.7.3
FROM:	<= 1.7.2
TOPIC:	BGP-related src primitives (ie. src_as_path) and JSON/Avro encoding
DESC:	Majority of these primitives were written in a 'src_' field, ie.
	src_as_path, contraddicting the prevailing name convention of other
	source/destination fields, ie. port_src, peer_ip_dst, etc.. These
	fields have now been aligned to the name convention - with src or
	dst qualifier at the end - and here is the full list of the fields
	modified: src_comms, src_ecomms, src_lcomms, src_med, src_as_path
	and src_local_pref; changed in: comms_src, ecomms_src, lcomms_src,
	med_src, as_path_src and local_pref_src respectively. 

TO:	>= 1.7.3
FROM:	<= 1.7.2
TOPIC:	BGP attribute "origin" encoding change
DESC:	The "origin" attribute was encded as an integer, until 1.7.2. To
	make output more human-friendly, the encoding has been changed to
	string in 1.7.3, where three single-letter values are defined:
	IGP (i), EGP (e) and Incomplete (u). Only release 1.7 releases
	starting with 1.7.7 do feature a tmp_bgp_daemon_origin_type_int
	true/false configuration knob to revert to the old encoding in
	order to allow for smoother pre-1.7.3 upgrades.

TO:	>= 1.7.2
FROM:	<= 1.7.1
TOPIC:  Obsoleted features
DESC:   Following is the list of features, knobs and plugins that are being
        discontinued with release 1.7.2:
	* pre_tag_map: matching on 'sampling_rate' is not supported anymore
	  since a sampling_rate primitive is available from version 0.14.2;
	  the 'return' feature to return matched data before completing the
	  map workflow has started the process of being obsolete (retired
	  from docs but still available).
	* nfacctd: collection, processing and replication of NetFlow v1, v7
	  and v8 has been discontinued. The focus will remain on popular v5,
	  v9 and IPFIX. 
	* tee_plugin: tee_dissect_send_full_pkt was allowing to specify if,
	  when dissecting replicated packets due to a complex pre_tag_map,
	  to send (or not) the full packet. Now, in a simplification move,
	  full packets are replicated only if no pre_tag_map is set or a
	  simple pre_tag_map is defined. In the context of replication, in
	  QUICKSTART is defined what is a simple and a complex pre_tag_map.
	* Kafka plugin: configuration directives kafka_avro_schema_topic and
	  kafka_avro_schema_refresh_time allow to produce an Avro schema at
	  regular time intervals to a topic; Confluent Platform supports an
	  Avro schema registry which is the clean/preferred way to handle the
	  distribution of schemas; as a result of this, the existing feature
	  is made legacy and will be discontinued at the next major release
	  and a new configuration directive kafka_avro_schema_registry is
	  introduced. 

TO:	>= 1.7.1
FROM:	<= 1.7.0
TOPIC:	nfprobe plugin, nfprobe_engine and NetFlow v9/IPFIX
DESC:	Until version 1.7.0 it was possible to (mistakenly) configure the
	NetFlow v9 SourceID field/IPFIX Observation Domain ID with the old
	NetFlow v5 jargon, ie. '1:1'. This is now threated as invalid and
	a positive 32-bit number, ie. '100000', is expected. If exporting
	NetFlow v5, nothing changed: the Engine ID/Engine Type input, ie.
	'1:1', is still valid and expected.

TO:	>= 1.7.1
FROM:	<= 1.7.0
TOPIC:	NetFlow/IPFIX/sFlow statistics via SIGUSR1
DESC:	If using this feature, please note that there has been a change in
	the output format. Mainly: 1) agent stats are now one per line, 2)
	the stats format has been formalised a bit and 3) +++/--- separators
	are sent at the beginning and at the end of the output (not anymore
	one per agent).

	The new output looks as follows:

	NOTICE ( default/core ): +++
	NOTICE ( default/core ): stats [0.0.0.0:2100] agent=X.X.X.X:0 time=1515772618 packets=1 bytes=496 seq_good=1 seq_jmp_fwd=0 seq_jmp_bck=0
	NOTICE ( default/core ): stats [0.0.0.0:2100] agent=Y.Y.Y.Y:0 time=1515772618 packets=2 bytes=992 seq_good=2 seq_jmp_fwd=0 seq_jmp_bck=0
	NOTICE ( default/core ): stats [0.0.0.0:2100] time=1515772618 discarded_packets=0
	NOTICE ( default/core ): --- 

	The old output was looking as follows:

	NOTICE ( default/core ): +++
	NOTICE ( default/core ): NetFlow statistics collector=0.0.0.0:2100 agent=X.X.X.X:0 (1515772618):
	NOTICE ( default/core ): Datagrams:       1
	NOTICE ( default/core ): Bytes:           496
	NOTICE ( default/core ): Good datagrams:  1
	NOTICE ( default/core ): Forward jumps:   0
	NOTICE ( default/core ): Backward jumps:  0
	NOTICE ( default/core ): ---
        NOTICE ( default/core ): +++
        NOTICE ( default/core ): NetFlow statistics collector=0.0.0.0:2100 agent=Y.Y.Y.Y:0 (1515772618):
        NOTICE ( default/core ): Datagrams:       2
        NOTICE ( default/core ): Bytes:           992
        NOTICE ( default/core ): Good datagrams:  2 
        NOTICE ( default/core ): Forward jumps:   0
        NOTICE ( default/core ): Backward jumps:  0
        NOTICE ( default/core ): ---
	NOTICE ( default/core ): +++
	NOTICE ( default/core ): Total bad NetFlow datagrams: 0 (1515772618)
	NOTICE ( default/core ): ---

TO:     >= 1.7.1
FROM:   <= 1.7.0
TOPIC:  Obsoleted features
DESC:   Following is the list of features, knobs and plugins that are being
        discontinued with release 1.7.1:
	* pkt_len_distrib primitive is discontinued due to apparent lack of
	  interest; the primitive was meant to bucket packet/flow/sample
	  lenghts in a distribution. Alternatively pre_tag_map can be used
	  to achieve the same. Buckets could be defined as follows:

	  set_tag=100	filter='ip[2:2] > 0 && ip[2:2] <= 512'
	  set_tag=200	filter='ip[2:2] > 512 && ip[2:2] <= 1024'
	  set_tag=300	filter='ip[2:2] > 1024 && ip[2:2] <= 9000'

	  Then the 'tag' primiive will carry the defined packet size bucket.
        * BGP daemon offline code, ie. bgp_daemon_offline_* directives, has
	  been deprecated in favor of alternative approaches, ie. BGP Looking
	  Glass (bgp_daemon_lg_*) and BGP Xconnects (bgp_daemon_xconnect_*).
	* removed --enable-threads / --disable-threads configure switch to
	  allow to compile pmacct even if no pthreads library is available.
	  From this release support for threads is mandatory. 

TO:	>= 1.7.1
FROM:	<= 1.7.0
TOPIC:	tag variables for dynamic Kafka topics and RabbitMQ routing keys.
DESC:	In an effort to harmonise variable names - and variable replacement
	functions - among the different plugins, $pre_tag has been renamed
	as $tag and $pre_tag2 as $tag2.  

TO:	>= 1.7.1
FROM:	<= 1.7.0
TOPIC:	allow file, ie. nfacctd_allow_file and equivalents.
DESC:	Definition of IP prefixes is now supported and hence the 0.0.0.0/0
	to allow all traffic in; whereas now an empty map rejects all input
	traffic. In previous releases an empty map would allow all traffic
	in instead and there was no way to simply reject all traffic. 

TO:	>= 1.7.1
FROM:	<= 1.7.0
TOPIC:	pmacct IMT client and timestamps
DESC:	Timestamps (ie. timestamp_start, timestamp_end, timestamp_arrival
	primitives) are now formatted in a rfc3339 compliant way, ie. if
	UTC timezone yyyy-MM-ddTHH:mm:ss(.ss)Z. This break compatibility
	as in previous releases timestamps had a more liberal formatting.
 
TO:	>= 1.7.0
FROM:	<= 1.6.2
TOPIC:	Obsoleted features
DESC:	Following is the list of features, knobs and plugins that are being
	discontinued with release 1.7:
	* MongoDB plugin is being discontinued since the old Mongo API is
	  not supported anymore and there has never been enough push from
	  the community to transition to the new/current API (which would
	  require a rewrite of most of the plugin)
	* Packet classification basing on the L7-filter project is being
	  discontinued (ie. 'classifiers' directive). This is being replaced
	  by an implementation basing on the nDPI project. As part of this
	  also the sql_aggressive_classification knob has been discontinued. 
	* tee_receiver was part of the original implementation of the tee
	  plugin, allowing to forward to a single target and hence requiring
	  multiple plugins instantiated, one per target. Since 0.14.3 this
	  directive was effectively outdated by tee_receivers. 
        * tmp_net_own_field knob was allowing to revert to backward compatible
          behaviour of IP prefixes (ie. src_net) being written in the same
          field as IP addresses (ie. src_host)
        * tmp_comms_same_field knob was allowing to revert to backward
          compatible behaviour of BGP communities (standard, extended) being
          writeen all in the same field.
        * plugin_pipe_amqp and plugin_pipe_kafka features were meant as an
          alternative to the homegrown queue solution for internal messaging,
          ie. passing data from the Core Process to Plugins, and are being
          discontinued. They are being replaced by a new implementation,
          plugin_pipe_zmq, basing on ZeroMQ.
	* plugin_pipe_backlog was allowing to keep an artificial backlog of
	  data in the Core Process so for plugins to maximise bypass poll()
	  syscalls in plugins. If home-grown queueing is found limiting,
	  instead of falling back to such strategies, ZeroMQ queueing should
	  be used. 

TO:	>= 1.7.0
FROM:	<= 1.6.2
TOPIC:	change to sql_num_hosts
DESC:	When sql_num_hosts is enabled and pmacct is not compiled with 
	--disable-ipv6, INET6_ATON() is now used for both IPv4 and IPv6 with
	MySQL and SQLite. Tables should be upgraded by changing columns from
	INT(4) to VARBINARY(16) and then converting data - for MySQL:

	ALTER TABLE table MODIFY COLUMN ip_src VARBINARY(16) NOT NULL;
	UPDATE table SET ip_src = INET6_ATON(INET_NTOA(ip_src))
	  WHERE INET_NTOA(ip_src) IS NOT NULL;

	This has performance implications which are mentioned in README.IPv6.
 
TO:	>= 1.6.2
FROM:	<= 1.6.1
TOPIC:	default plugin names
DESC:	Plugin names had to be unique per plugin type, meaning two plugins
	could be named "foobar" if they were of different type. Such behaviour
	has proven to lead to ambiguous scenarios and hence now plugin names
	must be globally unique. If not naming a plugin, its default name
	will now be "default_<plugin type>" instead of "default". Any piece
	of configuration that attaches directives to the "default" name, ie.
	relying on the assumption the plugin name defaults to "default", must
	be reviewed. 

TO:     >= 1.6.2
FROM:   <= 1.6.1
TOPIC:	print_time_roundoff configuration directive suppressed
DESC:	The directive was renamed print_history_roundoff for consistency with
	other plugins. print_time_roundoff was already removed from documents
	for the past 3+ years. 

TO:	>= 1.6.2
FROM:	<= 1.6.1
TOPIC:	sFlow probe (sfprobe plugin) and tags and class primitives
DESC:	Historically enterprise #8800 was squatted for the purpose of encoding
	pmacct-specific fields in sFlow, ie. tags and class. This never got
	changed when pmacct was assigned its own enterprise number (#43874) by
	IANA. In 1.6.2, these primitives are moved from #8800 to #43874 making
	older exports not compatible anymore. 
        
TO:	>= 1.6.1
FROM:	<= 1.6.0
TOPIC:	BGP communities and AS-PATH
DESC:	In pmacct 1.6.1 BGP communities and AS-PATH primitives (ie. std_comm,
	ext_comm, as_path, etc.) were moved from being fixed length to the
	variable-length framework, allowing for very long lists of communities
	or paths to be integrally represented. The IMT plugin is excluded from
	this work and these primitives will still be fixed length. Also, these
	primitives are not going to be supported anymore in the formatted output
	(but they will be no problem in the CSV, JSON and Avro output formats)
	of the print plugin. Btw, the formatted output support for the print
	plugin will be eventually discontinued in future so it is good to move
	away from it.

TO:	>= 1.6.1
FROM:	<= 1.6.0
TOPIC:	BGP extended communities (ext_comm, src_ext_comm primitives)
DESC:	Some legacy decision (..) made BGP standard and extended communities be
	written to the same field, mutual excluding each other. This behaviour
	has now been changed with each community type being written to the own
	field. For backward compatibility purposes a temporary config directive
	has been introduced, tmp_comms_same_field, that - if set to true - does
	enable the old behaviour. The config directive will be removed at the
	next major release.

TO:	>= 1.6.1
FROM:	<= 1.6.0
TOPIC:	print_markers
DESC:	In the print plugin, start marker is now printed also in the case where
	print_output_file_append is set to true; also, markers are printed as a
	JSON object, if output is set to JSON. 

TO:	>= 1.6.0
FROM:	<= 1.5.3
TOPIC:	uacctd switched from ULOG to NFLOG
DESC:	NFLOG supports both IPv4 and IPv6. While ULOG is still supported in
        recent kernels, NFLOG is supported since 2.6.14 and there is little
        point to support both - so a switch was made. The new daemon depends
        on the package libnetfilter-log-dev (in Debian/Ubuntu or equivalent
        in the prefered Linux distribution). For a quick test one can setup
	iptables to produce data in one of the following ways:

	* iptables -t mangle -I POSTROUTING -j NFLOG --nflog-group 4
	* iptables -t raw -I PREROUTING -j NFLOG --nflog-group 4

	And use the following command to collect data back:

	uacctd -c in_iface,out_iface,src_mac,dst_mac,src_host,dst_host,proto,src_port,dst_port -P print -g 4

TO:	>= 1.6.0
FROM:	<= 1.5.3
TOPIC:	build system refreshed
DESC:	autoconf and automake from early 2000 were being used to compile the
	build system until 1.5.3. This was for the sake of simplicity and
	robustness and, of course, came with drawbacks: somebody wanting to
	touch the build system should know which version of the tools to use,
	no leverage of the latest and greatest advancements made in the last
	one and half decades. The switch for should be almost transparent,
	the only impact being how to supply information in case the build
	system is unable to determine location of libraries (ie. via pkg-config
	and checking "typical" locations like /usr/local/lib): taking as an
	example PostgreSQL, before --with-pgsql-libs and --with-pgsql-includes
	were to be used to supply path to library and headers respectively;
	now environment variables PGSQL_LIBS and PGSQL_CFLAGS should be used
	instead for the same purpose, ie.:

	PGSQL_LIBS="-L/usr/local/postgresql/lib -lpq"
	PGSQL_CFLAGS="-I/usr/local/postgresql/include"
	./configure --enable-pgsql

TO:	>= 1.6.0
FROM:	<= 1.5.3
TOPIC:	nfacctd_disable_checks and sfacctd_disable_checks
DESC:	Default for this feature changed from false to true, ie. log warning
	messages for failing basic checks against incoming NetFlow/sFlow
	datagrams is disabled. For sequencing checks, the 'export_proto_seqno'
	primitive is recommended instead.

TO:	>= 1.6.0
FROM:	<= 1.5.3
TOPIC:	sql_recovery_logfile
DESC:	Feature removed from pmacct along with pmmyplay and pmpgplay logfile
	replay tools. 

TO:	>= 1.6.0
FROM:	<= 1.5.3
TOPIC:	MongoDB C legagy driver releases <= 0.8
DESC:	Support for MongoDB C legacy driver prior to 0.8 is dropped; in 0.8
	release, the most current version of the legacy driver, there was an
	impacting change of API; unfortunately in mongo.h the version was not
	updated and it looks the legacy driver is not maintained anymore (so
	no chance to have the nit fixed). The only way out seemed to default
	to the 0.8 behaviour, as that is the one currently being downloaded
	from GitHub by users. 

TO:     >= 1.6.0
FROM:   <= 1.5.3
TOPIC:  src_net and dst_net primitives
DESC:   Until 1.5.3 src_net and dst_net primitives value was written in the
        same field as src_host and dst_host - hence making the two sets mutual
        exclusive. This was found limiting by several users and, as a result of
        that, a separate field was added for storing networks (see "Increased
        memory usage by plugin caches" entry in this document). The use of such
        separate field had to be explicitely enabled by setting tmp_net_own_field
        configuration directive to true (by default set to false for backward
	compatibility); in version 1.6.0, tmp_net_own_field default value has
	now changed to true. tmp_net_own_field will be removed at the next
	major release.

TO:	>= 1.5.2
FROM:	<= 1.5.1
TOPIC:	--enable-ipv6 , IPv4-mapped IPv6 addresses & bindv6only
DESC:	Explicit support for IPv4-mapped IPv6 addresses was removed and now the
	bindv6only kind of behaviour is expected to be false (ie. both v4, via
	v4-mapped v6 addresses, and v6 addresses can connect to the v6 socket).
	On BSDs this is enforced in the code via a setsockopt() call; on Linux
	/proc/sys/net/ipv6/bindv6only is meant to enable/disable the feature.
	If binding to a "::" address (ie. no [sn]facctd_ip specified when pmacct
	is compiled with --enable-ipv6) no packets from IPv4 senders are not
	being received, then please check your bindv6only kernel setting. 

TO:	>= 1.5.2
FROM:	<= 1.5.1
TOPIC:	sql_history_since_epoch
DESC:	The effect of configuration directive sql_history_since_epoch has been
	ported to encompass any timestamp in pmacct, ie. timestamp_start and
	timestamp_end primitives, nfacctd_stitching, sfacctd counters filename,
	etc. The directive has hence been renamed timestamps_since_epoch. The
	old name, sql_history_since_epoch, has been removed from documentation
	but it is still going to be accepted in the configuration until the next
	major release for the sake of backwards compatibility.

TO:	>= 1.5.1
FROM:	<= 1.5.0
TOPIC:	Increased memory usage by plugin caches
DESC:	Source and destination IP prefixes aggregaton primitives, src_net and
	dst_net, now feature a separate field so to not be mutually exclusive
	with aggregation over IP addresses, ie. src_host and dst_host. In 1.5
	this can be optionally enabled by setting tmp_net_own_field to true;
	in later releases this behaviour will become default. The extra fields
	for IP prefixes do take additional memory in plugins cache - meaning 
	values for pre-allocated cache enries, ie. print_cache_entries, if
	configured to tight to available resources might generate SEGV and
	have to be reviewed downward.

TO:     >= 1.5.0
FROM:   <= 1.5.0rc3
TOPIC:  nfprobe plugin, NetFlow v9 export and flow timestamps
DESC:	timestamps for nfprobe plugin NetFlow v9 export are now absolute and
	in msecs, using field types #152 and #153. timestamps_secs can be set
	to true in order to revert to timestamps relative and in secs, using
	fields types #21 and #22.

TO:     >= 1.5.0
FROM:   <= 1.5.0rc3
TOPIC:  nfprobe plugin, NetFlow/IPFIX exports and tag, tag2 primitives
DESC:	tag and tag2 primitives can now be exported by nfprobe plugin only
	using IPFIX transport (nfprobe_version: 10). This is because, being
	custom pmacct field types, they have moved inside pmacct PEN for a
	cleaner solution (PENs not being supported by NetFlow v9). 

TO:	>= 1.5.0
FROM:	<= 1.5.0rc3
TOPIC:	NetFlow/IPFIX, print/AMQP/MongoDB plugins & time syncronization
DESC:	In 1.5.0 print/AMQP/MongoDB plugins are brought on par to SQL plugins
	by which flows/data with a future timestamp than the one currently
	being flushed is retained in the cache - to give further chances to
	in-memory data aggregation. This is intuitive, consistent behaviour
	but could happen time syncronization between collector and NetFlow/ 
	IPFIX agents was not an issue and suddenly it appears pmacct is not
	writing to the backend anymore. Solution is simply to sync all via
	NTP and use same timezone (recommended UTC for all).

TO:     >= 1.5.0rc3
FROM:   <= 1.5.0rc2
TOPIC:  nfacctd, sfacctd & plugin_pipe_size
DESC:	nfacctd_pipe_size and sfacctd_pipe_size configuration directives
	are being introduced in order to set the socket size between the
	daemons and the kernel. Until 1.5.0rc2 the same was accomplished,
	the dirty way, via existing plugin_pipe_size config directive when
	assigned to the core process. If relying on this trick on 1.5.0rc2
	and upgrading this can silently create packet loss on 1.5.0r3 and
	later (packet loss can be checked by veryfing that the counter
	showed by "netstat -s | grep Rcv" is not increasing).

TO:	>= 1.5.0rc3
FROM:	<= 1.5.0rc2
TOPIC:	MySQL plugin, additional libraries required when compiling
DESC:	MySQL 5.6 and later require linking against libstdc++ and librt. For
	this reason, when compiling MySQL plugin, it's now required that the
	development packages for these two libraries must be installed on the
	host system. Checks for this are introduced at configure script time.  
	It is not checked which MySQL version is installed so the requirement
	for these libraries is made retroactive.

TO:     >= 1.5.0rc3
FROM:   <= 1.5.0rc2
TOPIC:  SQL plugins, agent_id2 field
DESC:	Over the years, agent_id, agent_id2 fields were found confusing to
	store tag, tag2 primitives respectively. agent_id is now renamed 'tag'
	and backwards compatibility is preserved by issuing schema version #9.
	agent_id2 is not defined in any sql_table_schema instead and hence its
	renaming will be disruptive for existing deployments.

TO:     >= 1.5.0rc2
FROM:   <= 1.5.0rc1
TOPIC:  print plugin, dynamic file names and pointer to latest file
DESC:	Until 1.5.0rc1 pointer to latest file available was built as "<plugin
	name>-latest". Possibility to build variable spool directory structure
	and introduction of primitives-related variables, ie. $peer_src_ip, do
	phase-out the simple way of producing pointers, jeopardizing backward
	compatibility aswell. From 1.5.0rc2 a print_latest_file configuration
	directive allows to explicitely define pointer(s) to latest file(s):
	please refer to CONFIG-KEYS for more details about the feature. When
	upgrading, it is recommended to delete existing symlinks. 

TO:     >= 1.5.0rc2
FROM:   <= 1.5.0rc1
TOPIC:  print plugin, dynamic file names and time-related variables
DESC:	Time-related variables substitution is now based solely on the value of
	print_history. Previously, if print_history was not specified, this was
	based on the value of print_refresh_time. While this breaks backward-
	compatibility, it makes print plugin acting consistently to the rest of
	pmacct plugins. 

TO:     >= 1.5.0rc1
FROM:   <= 0.14.3
TOPIC:  print plugin, no entries to print_output_file
DESC:	In line with SQL plugins, in case there are no entries to account for the
	last print_refresh_time period, the purge function will not be invoked.
	As a result of that, if print_output_file contains time-based variables
	and if required to, output files will not be created anymore in case of
	no traffic to account for. Until 0.14.3, under same conditions, an empty
	output file (title only in case of formatted, CSV output) would have been
	printed out. 

TO:	>= 1.5.0rc1
FROM:	<= 0.14.3
TOPIC:	IPv6, peer_src_ip primitive, NetFlow exporter IP address
DESC:	Upon enabling IPv6 at compile time, via --enable-ipv6 switch, an IPv4
	NetFlow exporter IP address, ie. 10.0.0.1, was being written as IPv4-
        mapped IPv6 address, ie. ::ffff:10.0.0.1. This was causing confusion
	when composing maps, ie. the 'ip' field would change depending on whether
	IPv6 was enabled or not. To make maps consistent and simplify transitions
	to IPv6 compiled pmacct executables, IPv4-mapped IPv6 addresses are now
	internally translated to plain IPv4 ones.
 
TO:	>= 0.14.3
FROM:	<= 0.14.2
TOPIC:	networks_file & host aggregation primitives 
DESC:	In previous releases defining a networks_file in conjunction with host
	aggregation primitives would automatically work as a filter (ie. zero out
	hosts not included in the networks_file); whereas defining a networks_file
	in conjunction with net primitives would only work as a resolver. Now this
	behaviour has been streamlined by introducing a networks_file_filter true-
	false configuration directive to explicitely enable/disable the filtering
	feature (for both host and net primitives) on top of the resolver one. To
	summarize: if using a networks_file in conjunction with host aggregation
	primitives, and in order to keep the same behaviour while upgrading, a
	line should be added to the configuration: "networks_file_filter: true".

TO:     >= 0.14.3
FROM:   <= 0.14.2
TOPIC:	xlate_src and xlate_dst
DESC:	Feature has been obsoleted and replaced by proper aggregation primitives
	(nat_event, post_nat_*) to support NEL (NetFlow Event Logging) as currently
	implemented on Cisco ASR devices and to support CGNAT kind of scenarios.

TO:     >= 0.14.3
FROM:   <= 0.14.2
TOPIC:	nfacctd_sql_log
DESC:	Feature has been obsoleted and replaced by proper aggregation primitives
	(timestamp_start, timestamp_end) that effectively convert pmacct into a
	logger if enabled.

TO:     >= 0.14.0
FROM:   <= 0.14.0rc3
TOPIC:  peer_dst_ip
DESC:	The peer_dst_ip primitive is being attached to IP prefix resolution method
	(ie. as defined by nfacctd_net directive) from AS number resolution method
	in the past (ie. as defined by nfacctd_as_new directive).

TO:	>= 0.14.0
FROM:	<= 0.14.0rc3
TOPIC:	Fallback resolution of networks and ASNs (ie. nfacctd_net, nfacctd_as_new)
DESC:	Longest match wins has been introduced to select which route resolution
	method to use in fallback scenarios. For example up to 0.14.0rc3, a route
	advertised via BGP would have been winning over any more specific route
	learned via sFlow/NetFlow regardless. 

TO:     >= 0.14.0rc3
FROM:   <= 0.14.0rc2
TOPIC:  is_symmetric
DESC:   Support for is_symmetric aggregation primitive has been ceased due to lack
	of interest from the general community. 

TO:     >= 0.14.0rc3
FROM:   <= 0.14.0rc2
TOPIC:  peer_src_ip 
DESC:	peer_src_ip primitive must represent a reference (IP address, Agent ID) of
	the NetFlow or sFlow emitter for a certain flow. Due to previous work, this
	primitive was connected to the [ns]facctd_as_new mechanism which, if set to
	'bgp', was making it represent the IP address of a BGP peer instead. This is
	found not correct and hence peer_src_ip has now been disconnected from the
	[ns]facctd_as_new feature and always constitutes a reference to the NetFlow
	or sFlow emitter.

TO:	>= 0.14.0rc2
FROM:	<= 0.14.0rc1
TOPIC:	NetFlow v9 sampling
DESC:	Support for sampling in NetFlow v9 and IPFIX is elegant from an architecture
	point of view - but complex if compared to NetFlow v5 and sFlow for example.
	Such increased complexity lacking of proper framing by means of a supportive
	RFC exposes to bizzarre and creative implementations by vendors. 0.14.0rc2
	introduces fixes and workarounds to its sampled NetFlow v9 support in an
	effort to tackle specific but popular platforms among operators - and which
	can result in breaking some backward compatibility in this sense. 0.14.0rc2
	introduces a sampling_map feature, which although not rocket science from a
	concept point of view, it helps supporting sampled NetFlow v9 in heterogeneous
	network hardware environments at the cost of an extra static setting to care
	about; on the other hand it's also true sampling rates are often uniform and
	seldomly redefined in a production network. 

TO:     >= 0.12.1
FROM    <= 0.12.0
TOPIC:  Data source for ASNs must be explicitely defined
DESC:   data source for 'src_as' and 'dst_as' primitives for nfprobe and sfprobe
        plugins is now expected to be explicitely defined via the [ pmacctd_as |
        uacctd_as ] directive. All other plugins were already working like that.
        In terms of backward compatibility the only case affected is getting ASN
        values out of a Networks File: up to 0.12.0, it was sufficient to define
        a networks_file to implicitely use it.

TO:     >= 0.12.0rc1
FROM:   <= 0.11
TOPIC:  agent_id size and SQL table schemas
DESC:   With release 0.12, the agent_id field becomes 4-bytes large (from 2-bytes
        previously). SQL table schemas have been updated accordingly. If running
        a previous release and upgrading, you might incur into the risk that both
        Pre/Post-tagging infrastructures will accept values up to ~4M while the
        underlying SQL table schema is configured with a 2-bytes field. Solution
        is to run an "ALTER TABLE" statement to increase the field size during a
        maintenance window.

TO:     >= 0.12.0rc1
FROM:   <= 0.11
TOPIC:  nfprobe plugin: NetFlow v9 and 32-bit ASNs
DESC:   Release 0.12 introduces support for 32-bit ASNs in pmacct; things do not
        change in NetFlow v5 as if a 32-bit ASN is encountered, it is written as
        AS23456. In NetFlow v9, though, the source and destination AS fields are
        specified as 4 bytes long in the template. Given the template nature of
        NetFlow v9, this shouldn't pose a problem with 3rd party implementations
        but it's better to pay some extra attention while upgrading an existing
        installation.

TO:     >= 0.10.0
FROM:   <= 0.10.0rc3
TOPIC:  Configuration directives and command-line options
DESC:   In all previous releases, commandline options ( ie. -D -c ) were mutually
        exclusive with respect to configuration directives; now, they can cohexist
        and, more specifically, commandline options will override the content of
        the configuration file. This exposes to more interesting usages:

        shell> pmacctd -I <tracefile> -f <cfg>
        to launch pmacctd sharing an unique configuration file while reading data
        from different tcpdump/ethereal tracefiles among multiple runs.
 
TO:	>= 0.8.3 
FROM:	<= 0.8.2 
TOPIC:	Pre-Tagging, Post-Tagging
DESC:	In all previous releases, the 'pre_tag_map' and 'post_tag' directives were
        causing the captured traffic to be automatically tagged while forwarded to
	each active plugin; this behaviour can result in reduced flexibility; the
	0.8.3 release makes the two forementioned directives just to evaluate the
	tag to be assigned to captured traffic; a new 'aggregate' directive keyword
	- tag - causes the traffic to be marked (basing on the previous evaluation).
	So, a configuration like the following:

  	...
	pre_tag_map: /usr/local/pmacct/pre_tag.map
	aggregate[dummy]: src_host,dst_host,src_port,dst_port
	...

	Have to be rewritten the following way in order for the plugin 'dummy' to
	receive the tags:

	...
	pre_tag_map: /usr/local/pmacct/pre_tag.map
	aggregate[dummy]: tag,src_host,dst_host,src_port,dst_port
	... 

[EOF]