The Cacti Group | spine 1.2.31 -issue#365: Removed Backtrace Support due to lack of OS support -issue#366: Prevent polling from stopping when sysDescr OID returns noSuchObject -issue#370: Increase buffer sizes to handle large OID numbers -issue#373: Spine cannot generate SNMPv3 Ku for SHA-384/SHA-512 -issue#374: Improve depreciated DES support -issue#383: Improve Spine buffer handling #383 -issue#384: Buffer overflow in ping_icmp() causes Spine crash -issue: Clear up some minor memory leaks 1.2.30 -no changes 1.2.29 -no changes 1.2.28 -issue#350: When using Ping or SNMP Uptime, host is not always detected properly -issue#351: Add Ping Method where connection refused assumes host is up 1.2.27 -issue#349: Removal of AES Privacy Protocol was pre-mature 1.2.26 -issue#315: Errors cam be reported when uptime OID is not present -issue#317: MySQL reconnect option is depreciated -issue#321: Spine does not check a host with no poller items -issue#323: Poller may report the wrong number of devices polled -feature#329: Allow users to override the threads setting at the command line -feature#331: Allow spine to run in ping-only mode 1.2.25 -issue#234: Spine should see if script to be executed is executable -issue#291: Enhance number recognition -issue#294: When polling devices, sort by larger number of items first -issue#298: Log format may be corrupted when timeout occurs -issue#303: Compile warning appears due to GCC flag on RHEL7/RHEL8 -issue#304: Downed device detection only checks one of the two uptime OIDs -issue#305: Compile error appears due to execinfo.h on FreeBSD -issue#306: Bootstrap shell script contains some PHP cruft -issue#308: Padding is not always removed from the start of non-numeric strings -issue#311: Improve SNMP result handling for non-numeric results -issue#312: Further improve SNMP result handling for non-numeric results -issue#313: Remove check for the max_oids column which has been present since Cacti v1.0 -issue#314: Minimize Sorting when fetching poller records for maximum performance -feature#234: Spine should see if script to be executed is executable 1.2.24 -issue#296: When ignoring older OIDs, segmentation faults can be caused 1.2.23 -issue#287: Unexpected reindexing occurs when using uptime as the reindex method -issue#279: spine does not compile with SQL TLS deactivated -issue#298: Odd Cacti log format issue when script server timeout happened -feature#277: Spine should prevent the script server from connecting to remote when offline -feature#292: Improve Script Server Timeout Logging -feature#293: Add SQL_NO_CACHE to Spine Queries 1.2.22 -issue#265: Periodically spine crashes with sigabort due to MariaDB API buffer overruns -issue#267: FATAL: Spine Encountered a Segmentation Fault -issue#269: When polling time is exceed, spine does not always exit as expected -issue#270: Spine logging at `-V 5` includes an extra line feed -issue#271: Incorrect SNMP responses can cause spine to crash -issue#272: Properly handle devices that timeout responding to the Extended Uptime -issue#274: MariaDB can cause spine to abort prematurely despite error handling -feature#273: Spine should log the error time when exiting via signal 1.2.21 -issue#258: Disable DES if Net-SNMP doesn't have it 1.2.20 -issue#233: Add support for newer versions of MySQL/MariaDB -issue#238: When checking for uptime of device, don't assume a non-responce is always fatal -issue#240: When compiling, newer systems may complain about obsolete macros -issue#241: When querying devices, long descriptions can become truncated -issue#243: When compiling, FreeBSD may complain about missing backtrace functions -issue#244: When building queries, commands may become truncated -issue#247: When compiling with SNMPv3 support, SHA256 may incorrectly appear to be not supported -feature#252: Improve spine performance when only one snmp agent port is in use 1.2.19 -issue#172: Issues with polling loop may skip some datasources -issue#209: Ping no longer works due to hostname changes -issue#216: RRD steps are not always calculated correctly -issue#220: Unable to build when DES no longer supported -issue#221: IPv6 devices are not properly parsed -issue#226: Reduce a number of compiler warnings -issue#227: Compiler warnings due to lack of return in thread_mutex_trylock -issue#237: Spine will not look at non-timetics uptime when sysUpTimeInstance overflows -feature#235: Improve performance of Cacti poller on heavily loaded systems 1.2.18 -issue#207: Missing time parameter on FROM_UNIXTIME function 1.2.17 -issue#178: When building spine, errors with Percona libraries may appear by tersmitten -issue#184: When building spine, make process can report problems with backtrace -issue#186: When opening the database, spine can unexpectedly crash -issue#187: On large systems, too many connections can cause firewalls to assume a DDoS attack -issue#192: Spine does not properly use the mysql reconnect option -issue#193: When performing certain re-index operations, spine can unexpectedly crash -issue#194: When logging to standard output, log messaging is inconsistent -issue#195: When using newer gcc versions, spine is unable to compile properly -issue#197: Remote data collectors push too much data to main when performing diagnostics -issue#200: Make spine more responsive when remote connection is down -issue#202: MySQL SSL Key capability check does not work -issue#203: Backtracing support checks do not work -issue#204: MySQL Retry Count capability check always fails -feature: Make spine immune to DST changes -feature#181: Parse hostnames from tcp[6]:hostname:port [not fully implemented] -feature#188: On large systems, more concurrent script servers are needed 1.2.16 -issue: Some developer debug log messages falsely labeled as WARNINGS -issue#164: Remove the need of the dos2unix program -issue#171: Spine experiencing MySQL socket error 2002 under load -issue#175: Under heavy load MySQL/MariaDB return 2006 and 2013 errors on query -feature: Add backtrace output to stderr for signals -feature#176: Add Data Source turnaround time to debug output 1.2.15 -issue#166: Special characters may not always be ignored properly -issue#168: Correct issues with linking under OpenBSD 1.2.14 -issue#162: Spine not updating rrd_next_step 1.2.13 -issue#156: Compile on Cygwin is failing due to icmp6 headers missing 1.2.12 -issue#155: Failed host lookup causes spine to crash 1.2.11 -issue#122: Unable to compile spine on OpenBSD -issue#129: Repeated warnings due to 'Recache Event Detected for Device' -issue#134: No error is recorded to log file when database is incorrect -issue#144: MySQL retry logic is not working as expected -issue#149: When using IPv6 address, segmentation fault can occur due to incorrectly using IPv4 ping -issue#150: When polling results are null, segmentation errors may occur -issue#151: When collecting data, spine should be stripping alpha characters from output 1.2.10 -feature: release to match Cacti release 1.2.9 -issue#124: Improve logging of snmp multiple OID get function -issue#129: PCOMMAND Device[XXXX] WARNING: Recache Event Detected for Device -issue#130: Segmentation Fault triggered by a lot of errors -issue#131: Replace function gethostbyname with thread safe getaddrinfo -issue#139: removed strip_alpha for "Wrong-Type" SNMP-Answer 1.2.8 -issue#119: When calling bootstrap, an invalid error claiming config.h.in is missing can appear on first compile 1.2.7 -issue#113: When compiling under CentOS, 'm4' directory appears to be missing -issue#115: Perform proper initialization and teardown of sockets under Windows -issue#117: Spine configuration example shows remote database prefix for local database values 1.2.6 -issue#101: MySQL 8 has retired 'my_bool' type -issue#104: Spine should report which threads are outstanding -issue#105: Spine should include 'Process[]' in standard logging -issue#107: Log Buffer is too small to handle some output -issue#108: Once an SNMP timeout is reached, snmp_get does not honor include_host -issue#109: Initial declarations inside for loop cause compile errors -issue#110: When compiling, 'directive output may be truncated' messages can appear -issue#112: INSTALL instructions have typo in bootstrap command 1.2.5 -issue: Improve detection of MySQL vs. MariaDB client library requirement -issue#98: Checking type in poller.c causes compiler error 1.2.4 -issue#92: Add Data Source ID to "Empty result' error messages -issue#94: Increase default 'results_buffer' size to 2048 and MySQL max buffer to 131072 bytes 1.2.3 -issue#76: Spine unhandled exception signal -issue: More recent versions of MariaDB require different client library 1.2.2 -feature: release to match Cacti release 1.2.1 -feature: release to match Cacti release 1.2.0 -feature: Allow threads to be set a Data Collector level -issue#50: make fails on Debian 7 attempting to locate clock_gettime -issue#53: Improved escaping when updating database records -issue#54: Spine should always log if ICMP fails -issue#58: When updating snmp_sysLocation, text can become truncated -issue#63: Automatically reduce device threads when number of data sources does not require it 1.1.38 -feature: release to match Cacti release 1.1.37 -issue: Extra long snmp_sysObjectID's generate SQL errors 1.1.36 -issue#49: Re-Index update query failed due to lack of escaping 1.1.35 -issue#48: Availability check none triggers snmp system check that causes device to be ignored 1.1.34 -feature: release to match Cacti release 1.1.33 -feature: release to match Cacti release 1.1.32 -feature: release to match Cacti release 1.1.31 -issue#15: Spine cause CPU usage up to 100% -issue#32: Support noAuthNoPriv in Spine -issue#34: Re-introduce check for relocatable Net-SNMP and development files -issue#35: SNMP OID range checks cause some snmpget's to fail -issue#38: SNMP Ping timeouts not reported properly -issue#39: Spine should use same logging format as Cacti -issue#41: Spine cause CPU usage up to 100% -issue#42: Boost table keeps populating with boost disabled 1.1.30 -feature: release to match Cacti release 1.1.29 -issue#39: Have spine match cacti_log function 1.1.28 -feature: release to match Cacti release 1.1.27 -feature: release to match Cacti release 1.1.26 -feature: release to match Cacti release 1.1.25 -feature: release to match Cacti release 1.1.24 -feature: release to match Cacti release 1.1.23 -feature: release to match Cacti release 1.1.22 -feature: release to match Cacti release 1.1.21 -feature: release to match Cacti release 1.1.20 -issue#28: UDP and TCP ping issues 1.1.19 -feature: release to match Cacti release 1.1.18 -issue: Sort devices by polling time to allow long running devices to go first 1.1.17 -issue: Increase logging level for POPEN timeouts 1.1.16 -feature: release to match Cacti release 1.1.15 -feature: release to match Cacti release 1.1.14 -feature: release to match Cacti release 1.1.13 -issue: Script timeout should start after semaphore releases -issue: Don't include line breaks in Syslog 1.1.12 -feature: release to match Cacti release 1.1.11 -feature: release to match Cacti release 1.1.10 -feature: release to match Cacti release 1.1.9 -feature: release to match Cacti release 1.1.8 -issue#21: Errors when snmp_sysLocation contains a single quote 1.1.7 -feature: release to match Cacti release 1.1.6 -issue#15: Spine can cause high CPU utilization due to semaphore implementation -issue#11: Spine reports SNMP timeouts in some cases when gathering multiple OID in one request 1.1.4 -feature: release to match Cacti release 1.1.3 -feature: release to match Cacti release 1.1.2 -issue#13: Spine does not compile with --enable-lcap -issue#16: Spine not handling debug devices correctly 1.1.1 -feature: release to match Cacti release 1.1.0 -feature: release to match Cacti release 1.0.6 -feature: release to match Cacti release 1.0.5 -issue#7: remove database name not respected 1.0.4 -feature: release to match Cacti release 1.0.3 -issue#6: Spine accepts arbitrary polling modes 1.0.2 -feature: release to match Cacti release 1.0.1 -bug:With enough make jobs, help2man will fail if help2man is run before spine has been built 1.0.0 -feature: Track per host polling time -feature: Collect System Mib Information -feature: Place collection timing into host table -feature: Support snmpContextEngineID -feature: Change all references of Host to Device -bug:0001415: spine does not handle log_perror -bug#0002220: spine/cacti patch to allow SNMP_polling decices' hostnames like user@host:port -bug#0002234: Wrong alphanumeric value got from Spine SNMP poller -bug#0002651: False Positive Down status with snmpv3 -bug#0002674: Segmentation fault in Spine during polling host 0.8.8h -bug:0002659: wrong information about spine setup for icmps -bug: Removed requirement for building man page, deferred to 1.0 0.8.8g -bug:0002380: please build man page for spine -bug:0002610: Spine is trying to use mysql results after data being freed -bug:0002639: Spine does not respect snmp_retries from the database -bug:0002640: Spine crashes randomly while polling using multiple threads on one device -bug:0002549: Spine: hasCaps() always returns FALSE -bug:0002638: Spine: ping_udp() uses uninitialised socket timeout -bug:0002654: On RHEL 6.5, spine ./configure builds Makefile without crypto support 0.8.8f -bug:0002604: cacti-spine, configure script ignores LDFLAGS in certain situations 0.8.8e -bug#0002185: Multiple memory leaks found in spine -bug#0002398: Buffer overflow -bug#0002447: Building spine on Debian Jessie (testing) fails -bug#0002536: Spine segfaults when 18k+ hosts are to be pulled -bug#0002576: Call to a non-existent column of the database in poller.c -bug#0002578: spine doesn't compile on ubuntu 12.04 -bug#0002585: Spine: buffer overflow in logging when RESULTS_BUFFER>LOGSIZE can SEGV -bug#0002588: Spine "is_numeric" incorrectly returns TRUE for some hexadecimal values 0.8.8d -feature: add snmp system information gathering 0.8.8c -bug: remove obsolete reference causing pre-compile issues 0.8.8b -bug: set appropriate mysql 5.5+ timeouts 0.8.8a -bug#0002204: spine corrupted double-linked list error -bug#0002206: On multi-architecture Debian platforms spine fails to build -bug#0002205: Typo, twice, in spine.c: absense -> absence 0.8.8 -bug#0002185: Multiple memory leaks found in spine 0.8.7i -bug: Rework the way ping works, due to bad impact of changes in 0.8.7h -bug: ICMP ping not working on windows due to cygwin change -bug: Linux capabilities reported as on regardless of state -feature: Add a reconnect capability when the mysql server goes away 0.8.7h -bug#0001669: Problems with getting data from script using SPINE on Windows Server 2003 x32 -bug#0001829: Wrong string numerical value got from Spine SNMP poller -bug#0001899: spine treats undef replies as 0 -bug#0001968: Spine Error with Cisco FWSM firewalls -bug#0001974: PRIV_PROC_EXEC is required for Spine on Solaris 10 -bug#0002059: Spine doing too many (80+) connections when polling -bug: Net-snmp API issues cause spine crashes with some SNMP agents -bug: Host list not properly initialized -bug: Mutex locking issues cause deadlocks in Windows -bug: Escape windows type back slashes in scripts -bug: Use safer Non Net-SNMP string parsing calls for snmpget and snmpgetnext -bug: be smart for SNMP V3 and use AuthNoPriv when privacy passphrase is empty -bug: Do not fail on insert for a duplicate record -bug: Script server results with special characters can cause segfaults -feature#0001852: Another 10% speedup in spine -feature#0001878: Add Linux Capabilities to Spine -feature: Add additional ping methods to allow more precise snmp pings -feature: Clean up directory, update bootstrap 0.8.7g -feature: Multi threaded host polling 0.8.7f -bug#0001465: Don't check micro version when comparing installed net-snmp version to version used when building -bug#0001497: Prevent constant reindexes if snmp agent continually returns uptime of 0 -bug#0001510: Speedup spine resorting hosts by estimated polling time -bug#0001519: spine enhancement for parallel collection of data at host level -bug#0001541: Spine snmp bulk get incorrect assumes that the server SNMP Agent returns variables in same order -bug#0001551: Spine compile problem on AIX 5 -bug#0001563: Spine Compilation Problem on sql.c -bug#0001604: HEX Counter values enclosed in quotes not recognized as HEX -bug#0001657: Spine output shows drop of latest bit from hex value -bug#0001653: Transient ICMP Socket create failures causing spine to die -bug: When performing reindex checks, spine does not trim quotes from string values resulting in reindex faiures -bug: Add new option to enable Net-SNMP library vs. header versions -bug: ICMP fails on some platforms -bug: MySQL client auto reconnect default in MySQL 5.0.71 causes spine to crash -bug: Inconsistent behavior when the poller interval has not been set -bug: Invalid use of FDSETSIZE can cause issue with Script Server -bug: When performing reindex checks, spine does not trim quotes from string values resulting in reindex faiures -feature#0001629: Implement the snmp clientaddr option to bind spine to a specific local ip address 0.8.7e -bug: Not all configuration paths are searched for spine.conf -bug: If host has MAX OID's set to 0, timeouts occur -bug: Properly detect full Cygwin installs by finding sh.exe -bug: Don't perform TCP Ping retries in Windows. You cannot set socket send timeouts in Windows 0.8.7d -bug#0001310: Feature request for poller.php- single host polling -bug#0001313: Spine takes an excessive amount of time timing out failed tcp pings -bug#0001358: (cacti + spine) svn. Using snmp v3 constantly get "value: U", whereas snmpget returns correct value -bug#0001379: Ping.c might send more then retry_count icmp packages -bug#0001381: configure script distributed with cacti-spine-0.8.7c is broken -bug#0001420: PCOMMAND: Poller[0] Host[80] WARNING: Recache Event Detected for Host -bug: A host that for some reason can not initialize snmp can cause a segfault -bug: When using nifty popen, when the script times out, attempt to kill it -bug: Under some platforms tcp and udp ping unresponsive -bug: Availability messaging incorrect for snmpv3 hosts -bug: TCP Pings that receive a connection refused are good pings -bug: Avoid segmentation faults when Reindex table issues are found -bug: When performing snmp_get's, dont assume no error means non-null response -bug: Some snmp agents don't getnext on '.1' and need another oid '.1.3' -bug: Net-snmp versioning breaks snmpV3 -bug: Use system level snmp_retries variable instead of 3 -feature: Add one additional debug message to the poller code to track data sources polled -feature: Add more config paths for packagers -feature: Allow specification of an optional hostlist instead of first and last hostid's -feature: Integrate more tightly with boost by allowing writing directly to the boost output table -feature: Complete implementation poller_id field for people using a poller_id for polling. 0.8.7c -bug#0001022: strip_quotes incorrectly parsing script output -bug#0001097: Spine: Patch to allow unknown values including "Nan" and "U" without warnings -bug#0001101: spine hangs and times out -bug#0001130: ERROR: Spine Timed Out While Processing Hosts Internal -bug#0001147: spine doesn't trim white space from data template OIDs -bug#0001152: Running spine w/o setuid-root on Solaris -bug#0001186: clarify INSTALL documentation. -bug#0001231: Compiling Spine on Solaris 10 -bug#0001260: Deadlock in ping.c when resolving bad hostnames -bug#0001271: hex string to decimal conversion -bug#0001277: Spine reporting partial result or 'U' for SNMP value -bug: Correct a buffer overrun issue in poller.c -bug: If the script server goes away due to a syntax problem, it will not restart -bug: If the snmp session is not created spine times out -bug: If your snmp context changes between poller items, the context would not always change -feature: Use threadsafe gethostbyname_r -feature: Improve mysql logging -feature: Allow overriding the default results buffer size -feature: Allow overriding the default max mysql buffer (insert) size -feature: Allow overriding the maximum simultaneous scripts setting 0.8.7a -bug#0001054: Cactid/Spine installation instructions are missing dependencies -bug#0001056: Spine/Cactid configure reports incorrect URL for bug reporting -bug#0001071: When using Multiget OID's in Spine Errored OID's Cause Alignment Issues -bug#0001072: When the Last OID of a group is 'U', the host is repored down and remainder of OID's not polled -bug#0001097: Add patch to allow unknown values -bug#0001101: Fix ICMP hang while in a timed out state 0.8.7 -bug: Fix configure issues with hybrid 64 bit OS' -feature: Rename to spine -feature: Update copyright -feature: Add tab stops and autoindent to source code -feature: Make spike detection more generic -feature: Add snmpv3 support including contexts 0.8.6j -bug#0000879/0000872: cactid doesn't always report error messages before dying -bug#0000856: Slight documentation change -bug#0000861: Downed Host Detection on non-SNMP devices wrong -bug#0000886: SNMP Ping to Utilize The snmpgetnext instead of snmpget -bug#0000889: Use configure script's sysconfdir variable for placement of cactid.conf -bug#0000929: Spine with setuid root falls back to UDP after first host -bug#0000939: cactid use custom hostname, snmp_community,snmp_username and snmp_password in datasource -bug#0000940: Increasing BUFSIZE in cactid.h will still not allow longer output from scripts -bug#0000951: Spine SNMP polling does not strip enum values -bug#0000960: cactid truncating responses from scripts to 511 characters -bug#0000987: cactid can't use a socket unix -bug: By default, Cygwin can not handle large quantities of running scripts, therefore, limit to 20 -bug: Increase the number of file descriptors in Cygwin to 512, 64 does not work for large systems. -bug: If the system is out of resources when trying to fork(), retry three more times -bug: Use vfork instead of fork. It's more efficient with resources. -bug: Spine Script Server would not work with cgi version of php. Now it does. -bug: Changes in SNMP Community from poller item to poller item were not being considered. -bug: Restore meaningful notices when running from command line -feature: Allow users to use the traditional popen instead of nifty if required. -feature: Properly support INNODB database engine in MySQL 0.8.6i -bug: Only output as much as MySQL can handle in 1 insert -bug: Faulty Script Server shutdown was causing race condition -bug: Allow ICMP or UDP ping to work if host begins with either "TCP:" or "UDP:" -bug: Change assert logging to HIGH to more closely match cmd.php -bug: Allow Spine to make on 64bit x86 Platforms - thanks to Reuben Farrelly -bug: Changing syntax from "insert into to replace into" for reindex inserts -bug: Don't perform a reindex a check for a data query if you already have to reindex -bug: Make sure to log FATAL to stderr -bug: Do not re-index a host if the re-index value is a "No Such Instance" result. -bug#0000682: Allow storing of each hosts polling start time -bug#0000691: Spine fails with custom snmp port -bug#0000699: Tick counts not being detected -bug#0000706: Spine didn't start PHP scripts when host didn't answer SNMP (but icmp is ok) -bug#0000710: cactid.conf doesn't allow for no mysql password -bug#0000716: Don't attempt to save output for a host if there is none. -bug#0000717: Allow numeric double results to start with '+', '-', '.' -bug#0000724: cactid truncate script command when reindex -bug#0000736: Solaris 8 : Alarm Clock message -bug#0000745: cactid: search path for mysql and snmp does not include /usr/pkg -bug#0000748: cactid doesn't handle special charectors -bug#0000749: Spine 0.8.6g won't accept special characters from data template -bug#0000750: poller doesn't check for snmp if udp ping fails -bug#0000752: Problem with OID truncation with cactid -bug: add kstat checking for Net-SNMP v5.3.x -bug: get more intelegent. move cactid.conf.dist to /etc during install -feature: add Net-SNMP Library vs. Header check prior to execution -feature: enabled signal handling in cactid 0.8.6h -feature: Not released. Syncing with Cacti distribution again. 0.8.6g -bug#0000609: console "error" messages should go to stderr instead of stdout -bug#0000618/0000613: correct invalid assignment of session.community in snmp.c -bug#0000629: correct issues with invalid unknown reaults when reducing MAX OID get size -bug: correct issue where spike prevention logic was not being activated properly -bug: correct issue where host reindex was not operating for uptime goes backwards reindex method -bug: if sysUptime is queried for a host, only do it once -bug: certain SNMP string responses where not being handled properly due to net-snmp bug -bug: increase the allowable size of the logfile entry -bug: change behavior of availability so that if SNMP and ICMP/UDP are selected, only one has to succeed -bug: cleaned up implementation of multi SNMP OID get error checking logic -bug: don't attempt to allocate a zero item array when a host has no polling items -bug: correct string values returning as blank in *nix environments -bug: correct issue where invalid SNMP OID results would cascade to subseqent SNMP OID calls for a host -optimization: change logging logic to cleanup code and add varadic formatting -optimization: complete audit of code for portability complements of Steve @ http://www.unixwiz.net -optimization: documentation, documentation, documentation... -feature: added command line argument handling including database overrides complements of Steve @ http://www.unixwiz.net -feature: added multiple script server process feature to current version -feature: added support for MySQL 5.x -feature: added more verbose MySQL DEBUG logic 0.8.6f-1 -bug: Two memory leak fixes in poller.c -bug: Typo in the README-WINDOWS install file 0.8.6f -bug: allow a host to have more than 1 SNMP port defined for those odd devices that have them -bug: corrected an issue that caused infrequent segfaults when a host refused a ping request -bug: increased max_threads from 30 to 100 -bug: delayed killing php script server process by .5 seconds to allow MySQL connections to close -bug: place thread locks around unsafe functions to prevent possible segfaults -feature: implemented snmpv3 AuthNoPriv to align with Cacti 0.8.6 user interface -feature: improved performance of poller by the reduction of MySQL insert calls -feature: improved poller performance by grouping snmp get requests in a single call 0.8.6e -bug:#0000470: added -v, --version, -h, --help to command line options to make more GNU friendly -bug:#0000425: added ability to specify a socket file in *nix environments -bug: correct issues converting back slashes to forward slashes in arguments -bug: stop cactid from segfaulting when the database contains orphaned data source records -bug: re-implemented thread safe operations for non-linux systems -bug: allow cactid to poll from non-standard snmp port -bug: optimized some string handling routines -bug: changed snmp availability check to sysUptime from sysDesc -bug: limit the number of concurrent threads to 30 -bug: remove legacy cactid.conf file processing code -bug: add fault detection logic to mysql query calls to prevent segfaults -bug: cleaned up some unused variables -bug: add ability to detect errors and gracefully shut down php script_server -bug: remove trailing blanks and alpha characters from results -bug: clean up some string handling on script server results -bug: stop DEADLOCKS by not prematurely releasing mutex locks -bug: changed licensing to LGPL to accomodate linked libraries such as OpenSSL -bug: corrected issues where snprintf was taking 1 too many characters -bug: remove unneccessary Windows MACRO's from net-snmp initialization -feature: added seteuid functionality for unix and ICMP ping -feature: restart the script server after a timeout to prevent unexpected results -feature: only run the php script server if you have script server elements to process -feature: disable MIB loading to speed the launch of the net-snmp api -feature: change TREAD_SLEEP duration depending on number of processes to reduce CPU load -feature: allow the user to specify a script and script server timeout value from the user interface -feature: added polling for non-host specific data sources -feature: added data source logging to provide easier problem resolution -feature: added cmd.php like spike suppression -feature: added support for experimental multiple poller interval handling 0.8.6d -bug: increased timers for popen and script server to 20 seconds -bug: corrected segfault when polling a device with a null community string -bug: re-included ifdef for mysql thread protection for mysql 3.x 0.8.6c -bug: fixed integer/float conversion problem. -bug: increased the default script timeout from 5 to 10 seconds -bug: changed stdout format to reduce clutter on screen -bug: removed cygwin workaround that was causing SNMP instability -bug: increased both BUFSIZE and LOGSIZE to allow for larger output corner cases -bug: added mutex locking around MySQL calls and eliminated one unlock without a lock -feature: added connection timeout and retry logic -feature: added database port option in cactid.conf file. New parameter is DB_Port -feature: added some compiler directives for better compatibility with cygwin 0.8.6b -bug: fixed issue that prevented multi ds item RRD?s from updating -bug: fixed invalid string length calculation that was truncating output over 4 characters -bug: small log file reformatting ----------------------------------------------- Copyright (c) 2004-2024 - The Cacti Group, Inc.