Unnoc version 1.0.10.1 2007-07-11 Jason Schoonover For more details, take a look at doc/Changelog See the doc/UPGRADE file for upgrade instructions Unnoc version 1.0.10.1 New Features: - Added "about" link that will show what the current status of the daemon/cron wrapper is doing and show what hosts it is currently checking: display.php?about=1 - Updated host up/down code; If a host is already down, then it does not go through the full process of making sure the host is actually down; it just does a quick check one time through (instead of two times through) to check the host; this will cut down on the time spent on already downed hosts Bugfixes: - VI3 space in Virtual Machine name: If there was a space in the Virtual Machine Name, then the RRD would fail to be created; this has been fixed now, spaces (as well as other characters) are now escaped - There was a bug where if an RRD contained a space then it would fail, however this has been resolved now and updated - Fixed an issue, where under certain circumstances, with PHP 4.3.2, the math was done incorrectly on the VMware VI3 plugin, and completely inaccurate numbers were calculated. --------------------------------------------------------------------- ---- Unnoc version 1.0.10 (2007-06-27) --------------------------------------------------------------------- Details: New Features: - Added a show_down_interfaces option that allows a device to be configured to always show all interfaces. This is good for switches which will always show all graphs, regardless of the current state of the port. - Updated disk counters for both HR and UCD disks for server types (Linux, UNIX, etc); you can now manually specify the disk device if it is not picked up correctly. For instance, if you have a disk '/' and its device is /dev/md0, but Unnoc doesn't automatically pick it up, and you know that /dev/md0 is a combo of /dev/sda1 and /dev/sdb1, then you can specify like this: host { hostname = server1 type = server community = public disk { partition = / device = /dev/sda1, dev/sdb1 device_display = /dev/md0 } } And if the device I/O counters are available via SNMP, then it will graph and log the I/O. Otherwise it will manually pick it up. See the README.disks for more information. - Added Disk counters and Net counters to VMware VI3. This includes packet counters and throughput assessments for both ESX hosts and Virtual Machines. - Added more details to the Netapp page: including separating Aggregates and Tradition/Flexible volumes from the display page. - Added disk configuration for server and ms-server types per the unnoc.conf config file. If it is 'server' type, then it is used for disks that are larger than 2.2TB (or disks that just dont show up properly under the UCD-DISK table). For MS server the functionality was missing to be able to set disk warns / high usages. Also added ignore capability per disk as well. Example: host { hostname = server1 type = server community = public disktype = hr disk { partition = / warn_usage = 95% high_usage = 99% } disk { partition = /nfs/mount, /proc/bus/usb, /dev/shm ignore = 1 } host { server = server2 type = ms-server community = public disk { partition = C:, D: warn_usage = 80% } } See the documentation for more information Bug fixes: - VMware fix for Virtual Center servers managing ESX servers and standalone ESX servers coexisting; under some circumstances, if you viewed a VCMS link then it would display all ESX servers, instead of just the ESX servers that were managed by the VirtualCenter server. - VMware fix: CPU Usage fix; if an ESX server was using over 100% CPU, then the CPU Used column would display 0%, this is fixed now to display the proper value --------------------------------------------------------------------- ---- Unnoc version 1.0.9.1 (2007-06-07) --------------------------------------------------------------------- Unnoc 1.0.9.1 is a critical bug fix release. It addresses one major issue when a plugin (MySQL, OpenLDAP or PostgreSQL) was not responding correctly, it would alert until it was available, this is now fixed; also two other minor updates. Bug fixes: - APC PDU model numbers: updated APC PDU model numbers to pull the latest model number MIB from the APC - Updated VI3 links so that the wwwroot variable got set properly - All plugins (MySQL, OpenLDAP and PostgreSQL) were not getting updated properly in the database, so if one of them was down or not availble, it would send alerts until it was available; this has been fixed, it will only send alerts once --------------------------------------------------------------------- ---- Unnoc version 1.0.9 (2007-06-01) --------------------------------------------------------------------- Unnoc 1.0.9 is a very big release: there is now an unnoc daemon available. This is the first release that will offer both a cron method and a daemon method. The daemon method will help keep the load down on installations that have more than 200 hosts. In the process of writing the daemon, much of the code was optimized, including using pipes for RRDtool, this increased the performance of Unnoc quite a bit. Much of the code was cleaned up, as well as repeat code removed and consolidated. Details: New Features: - Daemon Mode. Unnoc can now run as a daemon. This helps with larger installations, especially where the load average is getting to be a bit high. See the README.daemon file for installation and configuration. I have provided a debian,ubuntu and fedora init.d/ script as well as a daemontools run.sh example file in the scripts/ directory. If anyone out there could do some init.d scripts for Suse, Gentoo, Solaris and FreeBSD that would help tremendously! - RRDTool Pipes. RRDtool is now run from a pipe within Unnoc, this means that unnoc will actually open up rrdtool only a few times and pipe all RRD graphs through this. It makes unnoc quite a bit quicker and decreases the load on the Unnoc server even further. - Optimized. Unnoc has been optimized considerably, consolidating alot of code down to the minimum. Many, many bugfixes with code overhaul. - APC battery failure alerts. On a APC Smart-UPS, if a battery fails then Unnoc will alert you and mark the APC appropriately. - Added host descriptions alert. If a host has a description tag, then unnoc will use this description tag, instead of the hostname, as the identifier when it sends out email alerts. To disable, use alert_with_description = 0. See README.config for more details - Added Disk I/O gathering for certain UCD/Net-SNMP supported devices. It doesnt work with all disks, for some reason it will only work with a few UCD-SNMP/Net-SNMP versions. - Added Unicast, Multicast, Discards and Error packet counters to all interfaces. - Simple DNS checking. This was long overdue. It has been missing from the beginning and I never got around to adding it. It is added now, if a host does not exist in DNS, then it will simply state this in the logs and move to the next host. Bug Fixes: - MS Server process case fix: under a few circumstances, two alerts would be sent out if the configured process name was not the same case as the name gathered from SNMP; this is fixed now, all processes are now case insensitive, regardless of configuration options or what SNMP says - VMware VI plugin: fixed a number of issues, including one where the VI3 plugin would not register hosts that were accessed by IP (not by hostname); also fixed ESX display issues: if an ESX server was not configured in SNMP, then it would not display under the VI3 page; this has been fixed - Uptime plugin: non-US timezones were not displaying correctly. This has been fixed now. - Major code change; many, many bugs fixed --------------------------------------------------------------------- ---- Unnoc version 1.0.8 (2007-04-30) --------------------------------------------------------------------- Unnoc 1.0.8 adds a few new features, but for the most part is just bugfixes. The most major new feature is process support in Windows, and CPU/Memory monitoring on a per process basis. This means that for every process that Unnoc is monitoring, it will graph the number of processes, the memory usage of the process, and the amount of CPU used by each process. Other new features include CPU usage for Cisco devices, more disk detail, new application plugins and APC battery alerts. Also, Unnoc now is not required to run as root. The ping times are a little longer (not quite as accurate), but you can run it now as an unprivileged user. New Features: - Unnoc is no longer required to run as root. See README.root for details. - Added Net::Ping::External support. If Net::Ping::External is available, it will be used instead of Net::Ping. This will use your OS version of ping, instead of the Perl module. This means that Unnoc is no longer required to run as root. See the README.root file for changing from root to non-priviledged user. - Added Plugin support for: - OpenLDAP (versions 2.1, 2.2 and 2.3) - MySQL (tested against 3.x, 4.x, and 5.x) - PostgreSQL (tested against 7.4 and 8.1) The plugin support for each application will graph and monitor statistics such as bytes in/out, number of operations, cache hits, connection counts, etc. See each of the respective documents for configuration information: README.openldap, README.mysql, README.postgresql. - Added a new updown type 'none'. This means that it will not let you know whether a host is available or unavailable. This is good for devices that you wish to monitor, but dont care if they go up or down, like for monitoring device attributes on employee's workstations. - Removed old ignore_hosts file, everything is now consolidated to the config file. Use 'ignore = 1' to ignore a host. - Added memory/CPU monitoring on a per process basis for *NIX servers (type 'server' in the configuration file); as a result, there is an upgrade script provided that will update all appropriate RRD files. See the UPGRADE file for more information. - Added process monitoring for MS servers; it is an unnoc.conf configuration option. This allows processes counts to be monitored, memory and cpu usage on a per process basis. Process count thresholds can be set per process (min/max allowed). Example: host { hostname = msserver1 community = public type = ms-server updown = ping proc { name = explorer.exe min = 1 max = 10 } proc { name = rdpclip.exe, winlogon.exe, services.exe min = 1 } ... } See README.hosts for more information - Added APC battery failure alerts; meaning that it will page/alert if an APC switches to batteries for any reason - Added more details for Disk Display for all devices that have disks - Added CPU Usage for Cisco Devices; any device now that is a Cisco product should have the type 'cisco-generic' instead of 'generic.' This will allow Unnoc to check a few extra MIB's for CPU percent usage. Only tested on single CPU, although multiple CPU devices should work as well. - Added more details for Interface graphs: added Interface Speed and added throughput display on the graph itself; some versions of Net-SNMP will report the wrong usage, particularly if you are using VLANs or are bonding interfaces together; I've noticed that it will default to 10mbit. You can correct this within the snmp configuration, by specifying speed. See the README.hosts for more information. Bug Fixes: - bug fix: alert/page settings were not being read properly for types ms-servers and barracuda spam firewalls; as a result, even though you specify no alert or no page, it would page/alert regardless; this is fixed now - minor bug fix for VI3. Added "0 VMs found" on ESX page and CVS page. Also added server statitsitcs for VMWare, if a server is not responding or if it is in mainteneance then it will say that it is - minor bug fix for process display; display had no lower limit set, so if a graph was alway at 1 process, for instance, then it will not display anything; now it will always start at 0 - minor bug fix for Barracuda Spam firewalls, temp_units were not displayed on zoom graphs, this is fixed now. --------------------------------------------------------------------- ---- Unnoc version 1.0.7 (2007-01-25) --------------------------------------------------------------------- This release has been long overdue. Over the past 10 months I have added all of the features listed below but did not get around to actually putting out a release until now. There are many new features below and many, many bugfixes. Details: New Features: - Added new type for Microsoft servers/workstations. Use the new type "ms-server", like this: host { hostname = server1 community = public type = ms-server updown = ping ... } This makes MS servers work properly and displays all of the stats correctly. - Added experimental support for external scripts that are called when a host is unavailable/available. See README.scripts for more information. - Added support for VMWare ESX 3 and Virtual Infrastructure 3 (VI3); Unnoc can now point to a VirtualCenter server or a standalone ESX server and read CPU and Memory statistics for each Virtual Machine that is connected to VI3. See README.vmware for more information. - Added support for Barracuda SPAM Firewalls; current support is for Barracuda SPAM Firewall 300 only (others have not been tested, although should work); Unnoc will graph the same data that is available from the admin interface on the Barracuda SPAM Firewall. See REAMDE.barracuda-spamfirewall for more information. - Added more statistics and graphs for Netapps: iscsi, fcp, nfs and cifs statistics; along with a per LUN read/write bytes and operations graph, also added entire disk statistics as well as CPU usage graphs; See the README.netapp docs for more information. - Added per disk partition limits for Netapps. Some versions of ONTAP will not set a limit for a disk partition, you can now add limits for your Netapp if you need them. See the README.netapp for more information. - Added logfiles and a configurable loglevel. This makes it possible to see exactly what unnoc is doing and get down to the gory details of why something is/isn't working. The log files are all located in the log/ directory: action.log -- Action Log command.log -- Misc command log mail.log -- Any email alert sent is logged here unnoc.log -- Log of devices going up/down, etc system_status.log -- All SS requests incoming/outgoing You can adjust the verbosity of all of the logs with the "loglevel" configuration option. - Added a configurable alert blackout period, which is a time range (HH:MM - HH:MM) where unnoc will not send any alerts out. This is very handy if you have a server that has a load spike every day, or any server that has the same symptom every day, but you don't want to be bothered by it. - Added the start of error checking. If you've ever received a "map_table" error, chances are it's related to DNS; this will at least catch that from now on, instead of just outputting a weird map_table message. - Added support for 3 phase APC Smart UPS. I've only tested this on a Smart-UPS VT 30kVA unit, however I'm pretty sure that it should work with other APC models (IXS and Symmetras). If anyone has one of these units out there and would like to test with it, please drop me a line and I will help out - Added support for paging groups and email groups. This will allow you to assign admins to particular servers so that not everybody will get paged or emailed for all servers. You can set it up to where admins will get paged for only servers that are specified for there alert_group; this makes unnoc much more flexible for alerting - Added a host status at the top of each display page, this will indicate why the host is marked as it is (it will say something like "1 alert" or "not responding" or "OK." This is something that should've been added a long time ago - For ping only sites, the default page will display the status of the host, and then it will display the latency graph. Bugfixes: - Fixed MS Windows collector. Previously it did not read anything exept the uptime. Now it will ready memory usage, interface statistics, CPU load, and disk usage. Still no processes or protocols plugin available. Tested against Windows XP Professional and Windows 2000 Server. - Fixed Aironet client count. Some older versions of IOS dont give the proper value for the number of clients; the fix counts them properly. - Fixed Aironet client display; under some specific circumstances, the client list would not get updated properly due to a misconfigured MySQL UPDATE statement; this is fixed now. Updated the same code for Apple Airports as well. --------------------------------------------------------------------- ---- Unnoc version 1.0.6.2 (2006-02-24) --------------------------------------------------------------------- Very big release, this is the last step in completely moving away from MRTG, which is no longer required. Config file structure has completely changed, many bug fixes and new features. Available in this release are tools to convert your existing MRTG installations to RRD, and a config file upgrader that will bring your config file up to the latest version. Many more bug fixes and new features added as well, details are below. - The config file has a new format, the format is on a per host basis. You can now specify almost every variable per host, so you can check different hosts at different intervals, and use different warnings/levels for different hosts. Many new configuration options per host, you can even group hosts together for displaying them easier. See the README.add_hosts and README.config for more information. - MRTG is no longer required! All plugins and hosts are now fully integrated with RRD, these include: Cisco Aironets, Apple Airports, APC Smart-UPS's, Environment Websensors, Netapp's and all hosts. I've also included an MRTG2RRD tool to convert all of your older MRTG graphs to RRD graphs. See the doc/UPGRADE file for more information on how to upgrade. - Href display functions were added. This means that manual links are no longer required on the index.php, all of the print_href_*() functions print the links. You can print by type, by hostname or you can group hosts together and print by group. See the README.index for more information. - Net::Telnet::Cisco is no longer required for Cisco Aironets, everything is done with 100% SNMP. The plugin also includes graphs/statistics on VLAN/SSID associations, and number of clients and traffic graphs. See the README.aironets for more information. - Apple Airport plugin updated; IP addresses now included (most of the time, see the docs for details on this), and graphs on each interface. Also the 'airports.php' file has been removed in favor of display.php. The way to view all clients attached to an airport is display.php?all=airports. See the README.airports for more information about the new features. - APC plugin was also updated, not only with RRD graphs, but much more information is available now via SNMP. See the README.apcs for more information. - Major speed increase in the SNMP libraries. My particular installation received a 40% speed increase (used to take 25 seconds to do everything, now it takes 12). This is due to a new algorithm that was implemented in all of the data collecting scripts. - Added minimum disk percentage support, you can now specify a minimum percentage in the snmpd.conf, instead of a value in kbytes. This makes configuring servers faster and easier, and much, much easier to read. - Added experimental procfix support. This is highly experimental but can be quite handy. It uses the procfix support built into snmpd.conf to try and correct processes that have problems. This means that if a process is either 'down' or 'too many/few', then a specified command will run to try and fix the problem, snmp server side. A read/write community is required and specific snmpd.conf configuration is also required. See README.procfix for more information - Added EM01 and APC to the system status plugin, previously the system status plugin would not warn about temperature, it does now. See the README.system-status for more information. - Added two options to the unnoc.conf file: alert, graph. You can now specify that you do not wish to receive alerts for a particular host (with the alert command) and you can also specify that you don't wish to graph hosts. Alert disabling can be handy if you want to monitor and graph user's workstations with graphs, for instance, , but don't care if they go down or get rebooted. - Updated the throughput plugin. All network interface now record this information, and you can view the throughput of any interface simply by clicking on the graph. The throughput can also be manually printed out via the display_throughput function. See the README.throughput file for more information. - Added more details to the RRD graphs, the X axis is much more useful now and easier to read, looks more like an MRTG graph (a time every 2 hours). - Many file's locations have changed in the lib/ directory, the upgrade.sh script will take care of all of it. See the doc/UPGRADE file for more informatoin - Many small changes, see the doc/Changelog for the gory details. Bugfixes: - Under some circumstances, if a ping host was marked as being down and then came back up, you would receive notifications of "host is back up" every x minutes and would continue to receive them until you reset it from the DB. It would never actually get marked as being up. This was due to the uptime column not existing, this is fixed. - Sometimes an uptime wouldn't get displayed if it was 0 days or less, this is fixed; it always gets displayed - Sometimes if a process was not "normal" it would not get updated with the latest information until it changed its state; this is fixed, it will get updated every time the script is called, regardless of what state it is in (good for updating information) - Development changes: - Many, many updates for development. The PHP functions file has been split up into 35 different files, and the amount of functions and code has nearly doubled. An entire directory structure now exists in the lib/ directory - Global variables are now defined in the lib/config-parse.pl and common/common.php file. They were removed out of unnoc.conf because they really shouldn't be touched. --------------------------------------------------------------------- ---- Unnoc version 1.0.5 (2005-10-20) --------------------------------------------------------------------- Big changes for 1.0.5. 1.0.5 is the first step in migrating away from MRTG and going strictly to RRD. Available in this release, are tools that will convert all of your existing MRTG logs/config files to RRD files for servers only (plugins and hosts are next, this is a start though). I've also added zoom features for all of the graphs, so that you can see closeups of everything. Also a few bug fixes, SNMP version 2, memory statistics and more features. Details are below Details: New Features: - Added misc server statistics for RRD, you can now monitor each network interface, memory usage, open TCP connections and CPU usage. These replace the MRTG graphs for each server, now you can remove MRTG from your configurations, and use customizable RRD graphs for everything. RRD graphs look much prettier, you can graph more than one thing at a time and make it look how you want it to look. I've included an mrtg2rrd.pl script that will convert all of the existing MRTG servers to RRDs. See the UPGRADE.rrdtool for more information - Added all colors to a config file called etc/graph_colors.conf, you can customize all of the graph's colors now, and make them whatever color you want them to be. See README.config for an explanation of all of the variables. - Added zoom RRD graph/image feature. You can now click on a graph of anything (like load average for instance) and it will redraw the graph without the limits. This is helpful, for instance, if you have a load average max set at 5.0 and a warning level of 3.0, but your load average never peaks above 0.5. The normal graph is almost unreadable, however if you click the graph it will enlarge it so that you can now read it. Also very useful for disk usage, you can graph exactly how fast you're eating (or freeing) up disk space. - Added support for SNMP version 2 and alternate SNMP port numbers. If you have any servers/hosts/firewalls that require SNMP version 2, you can now utilize them by specifying it in the unnoc.conf file. The syntax is host:port:version. See the new unnoc.conf file, or README.config. (This requires a change for throughput, see the UPGRADE file for more details) - Added a few MRTG adding scripts for plugin specific applications. We're going to be moving away from MRTG and going towards RRD, however in the mean time this will make it easier to add plugins (like aironets, apple airports, gateways, etc) to unnoc. Files are in the unnoc/mrtg/bin/ directory, they are add_aironet, add_gateway and add_aironet. Makes adding these to unnoc much, much easier. Included two MRTG template files in unnoc/mrtg/etc. - Added a memory usage chart to display_server_stats. Now you can see how much swap space/free space your server is really using just by looking at the stats page Bugfixes: - fixed the check-environ.pl hanging bug. Every now and then check-environ.pl would just hang for no apparent reason. This doesn't happen any more. - fixed division by zero division in functions.php for throughput plugin - Proper display of a server with no response, if it had been out of commission for longer than $no_response_tries Config changes: - Changed the throughput array to be compliant with the new format of host:port:version. Previously throughput was the only array where you could specify an alternate port number, now it's changed a bit. The upgrade-config.pl script takes care of all of that - Added graph colors in a separate file, so there is another configuration option for this. - Added a protocol array, for future protocol monitoring implementation - Removed mrtg_servers and www_mrtg_servers from unnoc.conf Development changes: - Changed all string 'NULL' occurrences in the database with real NULL references, and upgraded all scripts as a result of this. This requires an update to the DB. See the UPGRADE file on how to do this. --------------------------------------------------------------------- ---- Unnoc version 1.0.4 (2005-10-09) --------------------------------------------------------------------- Not many big changes like in 1.0.4, the only major change is that addition of graphing with RRD. unnoc will now graph all of your processes, load averages and disk usages. Very handy. Configuration is trivial. Details: New Features: - Added a config file upgrader, this makes upgrading a lot easier than before, the syntax is like this: ./upgrade-config.pl /webroot/ /location/of/unnoc-version/ This means that you don't have to worry about putting in any config file changes manually, just the script take care of it for you. See the README.config-upgrade for more information - Added RRD graphing support. This will graph all of your processes, load averages and disk usages (only bytes used currently). You can view the graphs by clicking on each of these in the status page. See the README.rrdtool for more information Bugfixes: - bug fix: added environ MRTG checking from unnoc-cron.pl. If you use the Eesensors websensor EM01, unnoc will now properly graph that. Previously it wasn't getting called from unnoc-cron.pl. It does now however. Development notes: Config changes: - added three new variables, rrd_enable, rrd_servers and rrd_devices. - removed variable throughput_time (this number is now calculated from cron_throughput) - remove variable version --------------------------------------------------------------------- ---- Unnoc version 1.0.3 (2005-09-24) --------------------------------------------------------------------- Big changes in 1.0.3, some good bug fixes and new features added. Details: New Features: - Added a crontab wrapper. You can remove all of the MRTG and unnoc related items from your /etc/crontab and just run unnoc-cron.pl every minute. It tries to intelligently poll the MRTG'd hosts to cut down on error messages, and everything is controlled from the unnoc.conf file. See README.cron-wrapper for more information - Added support for the Eesensors Environment monitor EM01. It will alert you if things get too hot, and you can display it very nicely on a php page. See README.em01 - Added support for the Apple Airport Extreme. The view is very similar to the Cisco Aironet output. See README.apple-airports - Added more documentation, including documentation explaining all parts of the unnoc.conf. See README.config - Added load and temperature alerts for APC's. The SNMP cards that you purchase for the APC's already have this functionality, however if you set the alerts lower for unnoc, then you have two lines of defense, also added humidity for the APC's. - Added uptime display for servers Bugfixes: - bug fix: updated the throughput plugin numbers to properly calculate the bandwidth usage. Previously it was updating with base 10 numbers, so I changed it to base 2 (1024 instead of 1000, for instance). The numbers are much more accurate now. - bug fix: fixed the load average code in check-stats.pl, it will now properly alert you and update load averages - bug fix: fixed the 'process not monitored' anymore bug, if a process wasn't showing up in SNMP anymore for some reason (some versions of net-snmp do this, depending on the load of the server) then unnoc would log it, email you that "the process is not being monitored anymore" and then remove it. Now it will just silently remove it if it goes away, and silently add it. - bug fix: updated the update-uptimes.pl script to not do a division by zero the first time it's run. Development notes: - updated the read_config() function in functions.php, it will now search for the etc/unnoc.conf file, so you don't have to specify it in the function call anymore (cleans up the index files). You can use the re-write-stats.pl file again if you want, otherwise it doesn't hurt to have the extra paths in there - added quite a few new php functions to display the server status. display_server_stats now displays uptime, and there's also four new functions -- status_*() in functions.php. You can use them to separate load average, process, host and disk usage problems on the main page. Basically you can have one dot per on the index.php, so that you can know exactly why an entry is read. I've included an example index.php (index-4.php) that shows this. - centralized a lot of functions in check-stats.pl / check-hosts.pl, cleaned up a lot of code Config changes: - aironets code was updated: it now pulls the description from the SNMP community instead of from the config file. So you'll have to update the unnoc.conf file with the community name of the aironets. See the doc/UPGRADE and upgrade/unnoc.1.0.3.conf file for more information - many other small bugfixes and cosmetic changes --------------------------------------------------------------------- ---- Unnoc version 1.0.2 (2005-09-12) --------------------------------------------------------------------- 1.0.2 has some major architecture changes for some of the perl scripts, specifically check-stats.pl which no longer updates index.html text files but instead writes to the mysql DB directly, the stats are then pulled via php function calls. This makes it quite a bit easier to centralize everything to the mysql DB. Details: - check-stats.pl no longer writes to index.html text files directly, it now writes to mysql DB directly and php function calls are used to view the status of a server - check-hosts.pl was rewritten to make more use of SQL code, its much less convoluted now and more straightforward (meaning less chance of there being a weird bug somewhere) - uptime actually does something now, it reads in the uptimes of all servers/devices that are monitored (vi SNMP) and will calculate the mean/average uptime based on all of the devices. Before it just displayed the uptime of the unnoc server - change to the aironets code. there's now a new script called check-aironets.pl, this basically has the exact same code as aironets.cgi, except it will write the output to a SQL table instead, so you can read it with a php function call. This means that you can put it at the top of your MRTG index page, so you can view exactly who is on along side the traffic graphs. - added a new variable to unnoc.conf file no_response_tries, this cuts down on flaky hardware and such. this means how many tries will unnoc accept 'no response' from server before alerting you. - also added another variable to unnoc.conf $wwwroot, this will fix relational paths such as /logos/ I've written a migration tool that will convert index.html files to index.php files (for servers, this isn't necessary for devices). It's called upgrade/re-write-stats.pl, the syntax is like this: upgrade/re-write-stats.pl /location/of/unnoc/ vim:tw=72:wm=1