|
Frequently Asked Questions + TroubleshootingABOUT QUESTIONS:
FAQ-ABO150 Which log format can AWStats analyze ? FAQ-ABO200 Which languages are available ? How to add my own language ? FAQ-ABO250 Can AWStats be integrated with PHP Nuke ? COMMON SETUP/USAGE QUESTIONS: Here, you can find the most common questions and answers about AWStats setup/usage process.
FAQ-COM050 What is the log size limit AWStats can analyze ? FAQ-COM090 Setup for FTP server log files (proftpd, vsftpd, ...). FAQ-COM100 Setup for MAIL log files (Postfix, Sendmail, QMail, MDaemon, Exchange). FAQ-COM110 Setup for MEDIA SERVER log files (Realmedia, Windows media, Darwin streaming server). FAQ-COM120 How to rotate my logs without loosing data. FAQ-COM130 How to run AWStats frequently ? FAQ-COM140 How to exclude my IP address (or whole subnet mask) from stats ? FAQ-COM145 How to use the Extra Sections features ? FAQ-COM150 Benchmark question. FAQ-COM200 How reverse DNS Lookup works, unresolved IP Addresses ? FAQ-COM250 Different results than other log analyzers (Analog, Webalizer, WUsage, wwwStats...). FAQ-COM300 Difference between local hour and AWStats reported hour. FAQ-COM350 How can I process old log file ? FAQ-COM360 How can I process several log files in one run ? FAQ-COM400 How can I update my statistics when I use a load balancing system that splits my logs ? FAQ-COM500 How can I reset all my statistics ? FAQ-COM600 How can I compile and build statistics on a daily basis only ? FAQ-COM700 Can I safely remove a line in AWStats history files (awstatsMMYYYY*.txt) ? ERRORS/TROUBLESHOOTING QUESTIONS: Here, you can find the most common questions and answers about errors or problems using AWStats.
FAQ-SET100 I see Perl script's source instead of its execution in my browser. FAQ-SET150 Error "...couldn't create/spawn child process..." with Apache for windows. FAQ-SET200 "Internal Error" or "Error 500" in a browser connecting to Apache. FAQ-SET210 "Internal Error" after a long time in my browser (See FAQ-COM100 "AWStats speed/timeout problems"). FAQ-SET220 Crash while running awstats.pl or page content only partialy loaded FAQ-SET250 Log format setup or errors. FAQ-SET270 Only corrupted/dropped records FAQ-SET280 Error "Not same number of records of...". FAQ-SET300 Error "Couldn't open file ..." FAQ-SET320 Error "Malformed UTF-8 character (unexpected..." FAQ-SET350 Empty or null statistics reported. FAQ-SET360 Statistics reported except for os, browsers, robots and keywords/keyphrases. FAQ-SET400 Pipe redirection to a file give me an empty file. FAQ-SET450 No pictures/graphics shown. FAQ-SET700 My visits are doubled for old month I migrated from 3.2 to 5.x FAQ-SET750 AWStats run out of memory during update process with cygwin Perl. FAQ-SET800 AWStats speed/timeout problems. SECURITY QUESTIONS: Here, you can find the common questions about security problems when setting or using AWStats.
FAQ-SEC150 How can I prevent some users to see statistics of other users ? FAQ-SEC200 How to manage log files (and statistics) corrupted by worms attacks like 'Code Red Virus like'. FAQ-ABO100 : WHICH SERVER LOG FILES OR OS ARE SUPPORTED ? AWStats can works with : Because AWStats is in Perl, it can works on all Operating Systems. Examples of used platforms (bold means 'tested by author', others were reported by AWStats users to work correctly) :
FAQ-ABO150 : WHICH LOG FORMAT CAN AWSTATS ANALYZE ? AWStats setup knows predefined log format you can use to make AWStats config easier. However, you can define your own log format, that's the reason why AWStats can analyze nearly all web, wap and proxy server log files. Some FTP servers log files, Syslog or mail logs can also be analyzed. The only requirement is "Your log file must contain required information". This is very short examples of possible log format: Apache common log format (see Note*), Apache combined log format (known as NCSA combined log format or XLF or ELF format), Any other personalized Apache log format, Any IIS log format (known as W3C format), Webstar native log format, Realmedia server, Windows Media Server, Darwin streaming server, ProFTPd server, vsFTPd server, Postfix, Sendmail, QMail, Mdaemon A lot of web/wap/proxy/streaming servers log format Note*: Apache common log format (AWStats can now analyze such log files but such log files does not contain all information AWStats is looking for. The problem is in the content, not in the format). I think analyzing common log files is not interesting because there is a lot of missing information: no way to filter robots, find search engines, keywords, os, browser. But a lot of users asked me for it, so AWStats support it. However, a lot of interesting advanced features can't work: browsers, os's, keywords, robot detection...). See also F.A.Q.: LOG FORMAT SETUP OR ERRORS . FAQ-ABO200 : WHICH LANGUAGES ARE AVAILABLE ? AWStats can make reports in 40 languages. This is a list of all of them, for last version, in alphabetical order (The code you can use for Lang and ShowFlagLinks parameter are the ISO-639-1 language codes):
But, you may find small documentation for other languages made by contributors on Documentation Contrib page. If your language is not in this list, you can translate it yourself. For this, find what is your 2 letter language code: here. Once, you get it, for example "gl" for Galician, copy the file awstats-en.txt into awstats-gl.txt, in langs directory and translate every sentences inside. You can do same for files inside tooltips_f, tooltips_m and tooltips_w sub-directories. Then send your translated file(s) to [email protected]. FAQ-ABO250 : CAN AWSTATS BE INTEGRATED WITH PHP NUKE ? I don't know any plan to make an Add-On for PHPNuke to include AWStats, for the moment. But this can change. You should ask to have a such Add-On to PHPNuke authors, and on PHPNuke forums. FAQ-COM025 : HOW TO USE AWSTATS WITH NO SERVER LOG FILE PROBLEM: I want to have AWStats statistics but i have no access to my server log file. SOLUTION: Because AWStats is a log analyzer, if you don't have any way to read your server log file, you have nothing to analyze and you should not be able to use AWStats. However, this is a trick that you can use to have a log file be build. You must add a tag to call a CGI script like pslogger into each of your web pages. This will allow you to have an artificial log file that can be analyzed by AWStats. You can find a Perl version of CGI pslogger enhanced by AWStats author here or a php version of CGI pslogger made by Florent CHANTRET here. FAQ-COM050 : WHAT IS THE LOG SIZE LIMIT AWSTATS CAN ANALYZE PROBLEM: I know I must run AWStats update process frequently on new log files, this means thoose files have a regular size, but for my first update, I want/need to run update process on old log files that are very large. Is there a limit on log file size AWStats can analyze ? SOLUTION: No. There is no limit in AWStats. This means you can use it on large log files (test were made on 10GB log files). However your system (Operating System or Perl version) might have a limit. For example, you can experience size limit errors on files larger than 2 or 4 GB. If limit is Perl only, try to use a Perl version compiled with "large file" option. If you can't find it nor build it, you can try to use a LogFile parameter that looks like this LogFile="cat /yourlogfilepath/yourlogfile |" instead of LogFile="/yourlogfilepath/yourlogfile" FAQ-COM090 : SETUP FOR FTP SERVER LOG FILES (proftpd, vsftpd, ...) PROBLEM: What do I have to do to use AWStats to analyze some FTP server log files ? SOLUTION: AWStats can be used with some FTP server log files. With ProFTPd: 1- Setup your server log file format: Modify the proftpd.conf file to add the following two lines :
To have the change effective, stop your server, remove old log file /var/log/xferlog and restart the server. Download a file by FTP and check that your new log file looks like this: [01/Jan/2001:21:49:57 +0200] ftp.server.com user RETR /home/fileiget.txt 226 1499 2- Then setup AWStats to analyze the FTP log file: Copy config file "awstats.model.conf" to "awstats.ftp.conf". Modify this new config file:
Now you can use AWStats as usual (run the update process and read statistics). With vsFTPd, or any FTP server that log with xferlog format: 1- Check your server log file format: Take a look at your FTP server log file. You must have a format that match the following example to use this FAQ :
2- Then setup AWStats to analyze the FTP log file: If your FTP log file format looks good, copy config file "awstats.model.conf" to "awstats.ftp.conf". Modify this new config file:
Now you can use AWStats as usual (run the update process and read statistics). FAQ-COM100 : SETUP FOR MAIL LOG FILES (Postfix, Sendmail, Qmail, MDaemon, Exchange...) PROBLEM: What do I have to do to use AWStats to analyze my mail log files ? SOLUTION: This tip works with AWStats 5.5 or higher. For Postfix, Sendmail, QMail or MDaemon log files You must setup AWStats to use a mail log file preprocessor (maillogconvert.pl is provided into AWStats tools directory, but you can use the one of your choice): For this, copy config "awstats.model.conf" file to "awstats.mail.conf". Modify this new config file:
Now you can use AWStats as usual (run the update process and read statistics). For Exchange log files Despite the high number of possible log format provided with Exchange, none of them is built enough seriously to offer an interseting analyze (missing informations, messy data, no id to join multiple records for same mail, etc...). For this reason, an "exact" log analysis is a joke with Exchange log files. However a little support is provided. In order to analyze Exchange traffic, you have to enable "Message Tracking" (see article http://support.microsoft.com/default.aspx?scid=kb;EN-US;246856). Then copy config awstats.model.conf file to "awstats.mail.conf". Modify this new config file:
Also don't forget that with Exchange, informations in a log analyses can't be exact. Do not send any questions or requests for using AWStats with Exchange, this is not a problem in AWStats and we have no time to support non opened products. If you want to have complete and accurate information with Exchange, forget using AWStats or use a more serious mail serveur (Postfix, Sendmail, QMail...) FAQ-COM110 : SETUP FOR A MEDIA SERVER (REALMEDIA, WINDOWS MEDIA SERVER, DARWIN STREAMING SERVER) PROBLEM: What do I have to do to use AWStats to analyze my Media Server log files. SOLUTION: For Realmedia Your log file will probably looks like this: 216.125.146.50 - - [16/Sep/2002:14:57:21 -0500] "GET cme/rhythmcity/rcitycaddy.rm?cloakport=8080,554,7070 RTSP/1.0" 200 6672 [Win95_4.0_6.0.9.374_play32_NS80_en-US_586] [80d280e1-c9ae-11d6-fa53-d52aaed98681] [UNKNOWN] 281712 141 3 0 0 494 Copy config awstats.model.conf file to "awstats.mediaserver.conf". Modify this new config file:
Now you can use AWStats as usual (run the update process and read statistics). For Windows Media Server / Darwin Streaming Server 1- If your Windows Media / Darwin streaming Server version allows it, setup your log format to write the following fields: c-ip date time cs-uri-stem c-starttime x-duration c-rate c-status c-playerid c-playerversion c-playerlanguage cs(User-Agent) cs(Referer) c-hostexe c-hostexever c-os c-osversion c-cpu filelength filesize avgbandwidth protocol transport audiocodec videocodec channelURL sc-bytes To have the change effective, stop your server, remove old log files and restart the server. Listen to streaming files and check that your new log file looks like this: 80.223.91.37 2002-10-08 14:18:58 mmst://mydomain.com/mystream 0 106 1 200 {F4A826EE-FA46-480F-A49B-76786320FC6B} 8.0.0.4477 fi-FI - - wmplayer.exe 8.0.0.4477 Windows_2000 5.1.0.2600 Pentium 0 0 20702 mms TCP Windows_Media_Audio_9 - - 277721 If your Windows Media/Darwin Streaming Server version does not allow to define your log format: Just follow instructions in step 2 directly but use the log format string found in first lines of your log files (Just after the "#Fields:" string) as value for AWStats LogFormat parameter. For example, you could have a LogFormat defined like this: LogFormat="c-ip date time c-dns cs-uri-stem c-starttime x-duration c-rate c-status c-playerid c-playerversion c-playerlanguage cs(User-Agent) cs(Referer) c-hostexe c-hostexever c-os c-osversion c-cpu filelength filesize avgbandwidth protocol transport audiocodec videocodec channelURL sc-bytes c-bytes s-pkts-sent c-pkts-received c-pkts-lost-client c-pkts-lost-net c-pkts-lost-cont-net c-resendreqs c-pkts-recovered-ECC c-pkts-recovered-resent c-buffercount c-totalbuffertime c-quality s-ip s-dns s-totalclients s-cpu-util" This means you don't use the AWStats tags but AWStats can often also understand all the IIS and/or Windows Media Server tags. 2- Then setup AWStats to analyze your Media Server log: Copy config awstats.model.conf file to "awstats.mediaserver.conf". Modify this new config file:
Now you can use AWStats as usual (run the update process and read statistics). FAQ-COM120 : HOW TO ROTATE MY LOGS WITHOUT LOOSING DATA PROBLEM: I want to archive/rotate my logs using my server system (for example logrotate) options or a third software (rotatelog, cronolog) but I don't want to lose any visits information during the rotate process. SOLUTION: The best way to do that on 'Linux like' OS is to use the linux built-in logrotate feature. You must edit the logrotate config file used for your web server log file (usually stored in /etc/logrotate.d directory) by adding the AWStats update process as a preprocessor command, like this example (bold lines are lines to add for having a prerotate process): /usr/local/apache/logs/*log { notifempty daily rotate 7 compress sharedscripts prerotate /usr/local/awstats/wwwroot/cgi-bin/awstats.pl -update -config=mydomainconfig endscript postrotate /usr/bin/killall -HUP httpd endscript } If using a such solution, this is sequential steps that happens:
The advantage of this solution is that it is a very common way of working, used by a lot of products, and easy to setup. You will notice that you can "lose" some hits: If you use the -HUP signal, you will only lose all hits that were written during D and E. Note that you will also break all requests still running at G. In the example, it's a 1 minute lost (for small or medium web sites, it will be less than few seconds), so this give you an error less than 0.07% (less for small web sites). This is not significant, above all for a "statistics" progam. If you use the -USR1 signal, you will not kill any request. But you will lose all hits that were wrote during D and E (like with -HUP) but also all hits that are still running after H (all very long request that requires several minutes to be served). If hit ends during I, it is wrote in a log file already analyzed, if hit ends at K, it is wrote nowhere. In the example, it's also a 0.07% error plus error for other not visible hits that were finished during I or K, but number of such hits should be very low since only hits that started before G and not finished after H are concerned. In most cases a hits needs only few milliseconds to be served so lost hits could be ignored. Note also that if you have x logrotate config files, with each of them a postrotate with a kill -HUP, you send a kill x times to your server process. So try to include several log files in same logrotate config file. You can have several awstats update command in the same prerotate section and you will send the -HUP only once, after all updates are finished. However, doing this, you will have a lap time between D and F (were some hits are lost) that will be higher. This is required for example if you use the cronolog or rotatelog tools to rotate your log files. For example, Apache users can setup their Apache httpd config file to write log file through a pipe to cronolog or rotatelog using Apache CustomLog directive: CustomLog "|/usr/sbin/cronolog [cronolog_options] /var/logs/access.%Y%m%d.log" combined If you use a such feature, you can't trigger AWStats update process to be ran just BEFORE the rotate is done, so you must run it AFTER the rotate process, so on the archived log file. To setup awstats to always point to last archive log file, you can use the 'tags' available for LogFile. The problem with that is that your data are refreshed only after a rotate has done. However, you will miss absolutely nothing (no hits) and your server processes are never killed. You run the awstats update process from you crontab frequently, every hour for example, and half and hour before the rotate has done. See next FAQ to know how to setup a scheduled job. Then, once the rotate has been done (by the logrotate or by a piped cronolog log file), and before the next scheduled awstats update process start, you run another update process on the archived log file using the -logfile option to force update on the archived log file and not the current log file defined in awstats config file. This will allow you to update the half hour missing, until the log rotate (AWStats will find the new lines). However don't forget that this particular update MUST be finished before the next croned update. FAQ-COM130 : HOW TO RUN AWSTATS UPDATE PROCESS FREQUENTLY PROBLEM: AWStats must be ran frequently to update statistics. How can I do this ? SOLUTION: A good way of working is to run the AWStats update process as a preprocessor of your log rotate process. See previous FAQ (FAQ-COM120) for this. But you can also run AWStats update process regularly by a scheduler: With Windows, you can use the internal task scheduler. The use of this tool is not an AWStats related problem, so please take a look at your Windows manual. Warning, if you use "awstats.pl -config=mysite -update" in your scheduled task, you might experience problem of failing task. Try this instead "C:\WINNT\system32\CMD.EXE /C C:\[awstats_path]\awstats.pl -config=mysite -update" or "C:\[perl_path]\perl.exe C:\[awstats_path]\awstats.pl -config=mysite -update" A lot of other scheduler (sharewares/freewares) are very good. With unix-like operating systems, you can use the "crontab". This is examples of lines you can add in the cron file (see your unix reference manual for cron) : To run update every day at 03:50, use : 50 3 * * * /usr/local/awstats/wwwroot/cgi-bin/awstats.pl -config=mysite -update >/dev/null To run update every hour, use : 0 * * * * /usr/local/awstats/wwwroot/cgi-bin/awstats.pl -config=mysite -update >/dev/null FAQ-COM140 : HOW CAN I EXCLUDE MY IP ADDRESS (OR WHOLE SUBNET MASK) FROM STATS ? PROBLEM: I don't want to see my own IP address in the stats or I want to exclude counting visits from a whole subnet. SOLUTION: You must edit the config file to change the SkipHosts parameter. For example, to exclude: FAQ-COM145 : HOW TO USE THE EXTRA SECTIONS FEATURES ? PROBLEM: I want to build personalized reports not provided in default AWStats reports. How can I setup the Extra Sections parameters in my AWStats config file to do so ? SOLUTION: Take a look at the Using AWStats Extra Sections features FAQ-COM150 : BENCHMARK / FREQUENCY TO LAUNCH AWSTATS TO UPDATE STATISTICS PROBLEM: What is AWStats speed ? What is the frequency to launch AWStats process to update my statistics ? SOLUTION: All benchmarks information and advice on frequency for update process are related into the Benchmark page. FAQ-COM200 : HOW REVERSE DNS LOOKUP WORKS, UNRESOLVED IP ADDRESSES PROBLEM: The reported page AWStats shows me has no hostnames, only IP addresses, countries reported are all "unknown". SOLUTION: When AWStats find an IP address in your log file, it tries a reverse DNS lookup to find the hostname and domain if the DNSLookup parameter, in your AWStats config file, is DNSLookup=1 (Default value). So, first, check if you have the good value. The DNSLookup=0 must be used only if your log file contains already resolved IP address. For example, when you set up Apache with the HostNameLookups=on directive. When you ask your web server to make itself the reverse DNS lookup to log hostname instead of IP address, you will still find some IP addresses in your log file because the reverse DNS lookup is not always possible. But if your web server fails in it, AWStats will also fails (All reverse DNS lookups use the same system API). So to avoid AWStats to make an already done lookup (with success or not), you can set DNSLookup=0 in AWStats config file. If you prefer, you can make the reverse DNS lookup on a log file before running your log analyzer (If you only need to convert a logfile with IP Addresses into a logfile with resolved hostnames). You can use for this logresolvemerge tool provided with AWStats distribution (This tools is an improved version of logresolve provided with Apache). FAQ-COM250 : DIFFERENT RESULTS THAN OTHER ANALYZER PROBLEM: I also use Webalizer, Analog (or another log analyzer) and it doesn't report the same results than AWStats. Why ? SOLUTION: If you compare AWStats results with an other log file analyzer, you will found some differences, sometimes very important. In fact, all analyzer (even AWStats) make "over reporting" because of the problem of proxy-servers and robots. However AWStats is one of the most accurate and its "over reporting" is very low where all other analyzers, even the most famous, have a VERY HIGH error rate (10% to 200% more than reality !). This is the most important reasons why you can find important differences: There is also other reasons, however those points explains only small differences: AWStats has a larger browsers, os', search engines and robots database, so reports concerning this are more accurate. AWStats has url syntax rules to find keywords or keyphrases used to find your site, but AWStats has also an algorithm to detect keywords of unknown search engines with unknown url syntax rule. AWStats does not count twice (by default) redirects made by rewrite rules that makes two hits into log files but that are only one page "viewed". Etc... If you want to check how serious is your log analyzer, try to parse the following log file. It's a very common log file but results will show you how bad are most log analyzers (above all commercial products):
This is what you should find: 6 true human visits 5 different true visitors 1 bot visit 1 worm attack The entry pages for true visits should be "/" (even for 80.8.55.1) or "/cgi-bin/order.cgi" but nothing else. Note: I did not find any commercial log analyzer that can deal such a common log file correctly, so if you find, let me know ! FAQ-COM300 : DIFFERENCE BETWEEN LOCAL HOURS AND AWSTATS REPORTED HOURS PROBLEM: I use IIS and there's a difference between local hour and AWStats reported hour. For example I made a hit on a page at 4:00 and AWStats report I hit it at 2:00. SOLUTION: This is not a problem of time in your local client host. AWStats use only time reported in logs by your server and all time are related to server hour. The problem is that IIS in some foreign versions puts GMT time in its log file (and not local time). So, you have also GMT time in your statistics. You can wait that Microsoft change this in next IIS versions. However, Microsoft sheet Q271196 "IIS Log File Entries Have the Incorrect Date and Time Stamp" says: The selected log file format is the W3C Extended Log File Format. The extended log file format is defined in the W3C Working Draft WD-logfile-960323 specification by Phillip M. Hallam-Baker and Brian Behlendorf. This document defines the Date and Time files to always be in GMT. This behavior is by design. So this means this way of working might never be changed, so another chance is to use the AWStats plugin 'timezone'. Warning, this plugin need the perl module Time::Local and it reduces seriously AWStats speed. To enable the plugin, uncomment the following line in your config file. LoadPlugin="timezone TZ" where TZ is value of your signed timezone (+2 for Paris, -8 for ...) FAQ-COM350 : HOW CAN I PROCESS AN OLD LOG FILE ? PROBLEM: I want to process an old log file to include its data in my AWStats reports. SOLUTION: You must change your LogFile parameter to point to the old log file and run the update (or use the -LogFile option on command line to overwrite LogFile parameter). The update process can only accept files in chronological order for a particular month, so if you have already processed a recent file and forgot to run update on a log file that contains older data, you must before reset all your statistics (see FAQ-COM500) and restart all the update processes for all past log files and in chronological order. However, there is a "tip" that allow you to rebuild only the month were you missed data: Imagine we are on 5th of July 2003, all your statistics are up to date except for the 10th of April 2003 (you forgot to run the update process for this day, so there is no visit for this day). You can : - Reset the statistics for April only (this means remove the file awstats042003.[config.]txt as explained in FAQ-COM500), - Move the statistics history files for month after April (file awstats052003.[config.]txt, awstats062003.[config.]txt,...) into a temp directory (so that it is no more in DirData directory as if they were deleted). - Run update process on all log files for April (in chronological order). AWStats does not complain about "too old record" because there is no history files in DirData directory that contains compiled data more recent than records into log you process. - Moved back the month history files you saved into your DirData directory. Your statistics are up to date and the missing days are no more missing. FAQ-COM360 : HOW CAN I PROCESS SEVERAL LOG FILES IN ONE RUN ? PROBLEM: How can I update my statistics for several log file, in one run ? SOLUTION: A solution should be to setup your config file with something like: LogFile=mylog*.log However, with such a syntax, AWStats can't know in wich order processing log files (wich log file is the first, next or last). So to work like this you must use the following syntax: LogFile="/pathto/logresolvemerge.pl mylog*.log |" Logresolvemerge is a tool provided with AWStats (in tools directory) that merge several log files on the fly sending line by line always the older record from a list of several log files. Using such a tool as a pipe source for AWStats LogFile parameter is a very good solution because, it allows you to merge log files whatever their size with no memory use, no hard disk use (no temporary files built), it is fast, it prevents you from a bad order if your log files are not correctly ordered, etc... This tool can also be used to process log files from load balanced systems (see FAQ-COM400) FAQ-COM400 : HOW CAN I UPDATE MY STATISTICS WHEN I USE A LOAD BALANCING SYSTEM THAT SPLITS MY LOGS ? PROBLEM: How can I update my statistics when i use a load balancing system that split my logs ? SOLUTION: First solution is to merge all split log files resulted from all your load balanced servers into one. For this, you can use the logresolvemerge tool provided with AWStats : logresolvemerge.pl file1.log file2.log ... filen.log > newfiletoprocess.log And setup the LogFile parameter in your config file to process the newfiletoprocess.log file or use the -LogFile command line option to overwrite LogFile value. As an other solution, if you miss disk space, or to save time, you can ask logresolvemerge to merge log files on the fly during the AWStats update process. For this, you can use the following syntax in your AWStats config file: LogFile="/pathto/logresolvemerge.pl file*.log |" See also FAQ-COM360 for explanation on logresolvemerge use. FAQ-COM500 : HOW CAN I RESET ALL MY STATISTICS ? PROBLEM: I want to reset all my statistics to restart the update process from the beginning. SOLUTION: All analyzed data are stored by AWStats in history files called awstatsMMYYYY.[config.]txt (one file each month). You will find those files in directory defined by DirData parameter (same directory than awstats.pl by default). To reset all your statistics, just delete all files awstatsMMYYYY.txt To reset all your statistics built for a particular config file, just delete all files awstatsMMYYYY.myconfig.txt Warning, if you delete those data files, you won't be able to recover your stats back, unless you kept old log files somewhere. You will have to process all past log files (in chronological order) to get your statistics back. FAQ-COM600 : HOW CAN I COMPILE AND BUILD STATISTICS ON A DAILY BASIS ONLY ? PROBLEM: How can I compile and build statistics on a daily basis. I mean i want to have a full report with all charts with data for a particular day only and want one report for each day of month. SOLUTION: This is an non documented and not supported trick, as this is not the standard way of working: First, run the update process at midnight (or on a log file that was rotated at midnight so that it contains only data for this particular day (you can choose another hour in night if you want to have days that "start" at an different hour). Once the update process has been ran, MOVE (and not copy) the history file built by AWStats. For example on Unix like systems: mv mydirdata/awstatsMMYYYY.mydomain.txt mydirdate/awstatsDDMMYYYY.mydomain.txt Note that the name has been changed by adding the day. Repeat this each day after the update process. With this you will have one history file for each day. You can then see full stats for a particular day by adding the non documented parameter -day=DD on command line (with others like -month=MM and -year=YYYY). If ran from a browser you can also add &day=DD on URL. However, if you have full day by day statistics, you don't have anymore statistics for full month, except if you create a second config file that whose history files would not be moved. FAQ-COM700 : CAN I SAFELY REMOVE A LINE IN HISTORY FILES (awstatsMMYYYY*.txt) ? PROBLEM: After processing a log file I want to change my statistics without running AWStats update process but changing directly data in AWStats historical database files. SOLUTION: If you remove a lines starting with "BEGIN_" or "END_", AWStats will find your file "corrupted" so you must not change those two kinds of lines. You can change, add or remove any line that is in any sections but if you do this, you must also update the MAP section (lines between BEGIN_MAP and END_MAP) because this section contains the offset in file of each other sections for direct I/O access. If history file is the last one, you can easily do that by removing completely the MAP section and run an update process. Like that AWStats will rewrite the history file and the MAP section will be rewritten (MAP section is not read by update process, only written). You do this at your own risk. The main risk is that some charts will report wrong values or be unavailable. FAQ-SET050 : ERROR "MISSING $ ON LOOP VARIABLE ..." PROBLEM: When I run awstats.pl from command line, I get: "Missing $ on loop variable at awstats.pl line xxx" SOLUTION: Problem is in your Perl interpreter. Try to install or reinstall a more recent/stable Perl interpreter. You can get new Perl version at ActivePerl (Win32) or Perl.com (Unix/Linux/Other). FAQ-SET100 : I SEE PERL SCRIPT'S SOURCE INSTEAD OF ITS EXECUTION PROBLEM: When I try to execute the Perl script through the web server, I see the Perl script's source instead of the HTML result page of its execution ! SOLUTION: This is not a problem of AWStats but a problem in your web server setup. awstats.pl file must be in a directory defined in your web server to be a "cgi" directory, this means, a directory configured in your web server to contain "executable" files and not to documents files. You have to read your web server manual to know how to setup a directory to be an "executable cgi" directory (With IIS, you have some checkbox to check in directory properties, with Apache you have to use the "ExecCGI" option in the directory "Directive"). FAQ-SET150 : INTERNAL ERROR 500 IN MY BROWSER FAQ-SET200 : ERROR "... COULDN'T CREATE/SPAWN CHILD PROCESS..." PROBLEM: AWStats seems to run fine at the command prompt but when ran as a CGI from a browser, I get an "Internal Error 500". I also also might have the following message in my Apache error log file (or in browser with Apache 2.0+): ...couldn't create/spawn child process: c:/mywebroot/cgi-bin/awstats.pl SOLUTION: First, try to run awstats.pl from command line to see if file is correct. If you get some syntax errors and use a Unix like OS, check if your file is a Unix like text file (This means each line end with a LF char and not a CR+LF char). If awstats.pl file runs correctly from command line, this is probably because your web server is not able to known how to run perl scripts. This problem can occur with Apache web servers with no internal Perl interpreter (mod_perl not active). To solve this, you must tell Apache where is your external Perl interpreter. For this, you have 2 solutions: 1) Add the following directive in your Apache httpd.conf config (or remove the # to uncomment it if line is already available) ScriptInterpreterSource registry Then restart Apache. This will tell Apache to look into the registry to find the program associated to .pl extension. 2) Other solution (not necessary if first solution works): Change the first line of awstats.pl file with the full path of your Perl interpreter. Example with Windows OS and ActivePerl Perl interpreter (installed in C:\Program Files\ActiveState\ActivePerl), you must change the first line of awstats.pl file with: #!c:/program files/activestate/activeperl/bin/perl FAQ-SET220 : CRASH WHILE RUNNING AWSTATS.PL OR PAGE CONTENT ONLY PARTIALY LOADED ON WINDOWS XP PROBLEM: Sometimes my browser (Most often IE6) crash while running awstats.pl with some AWStats configuration. With some other versions or browsers, page content is partialy loaded. SOLUTION: Problem was with WinXP and WinXPpro as documented at MS site Q317949; "Socket Sharing Creates Data Loss When Listen and Accept Occur on Different Processes" Result was that MSIE would crash or display nothing. Netscape and Opera handled the socket better but displayed the pages partially. The effect of the bug was more prononced as the page contents increased (above 30k). http://support.microsoft.com/default.aspx?scid=kb;EN-US;q317949 And also at Apache.org http://www.apache.org/dist/httpd/binaries/win32/ MS produced a Hotfix which is now included in SP1. But the best solution is to use a better web browser. Take a look at Firefox, one of the best and most popular web browser. FAQ-SET250 : LOG FORMAT SETUP OR ERRORS PROBLEM: Which value do I have to put in the LogFormat parameter to make AWStats working with my log file format ? SOLUTION: The AWStats config file give you all possible values for LogFormat parameter. To help you, this is some common cases of log file format, and the corresponding value for LogFormat you must use in your AWStats config file:
virtualserver1 62.161.78.73 - - [dd/mmm/yyyy:hh:mm:ss +0x00] "GET /page.html HTTP/1.1" 200 1234 "http://www.from.com/from.htm" "Mozilla/4.0 (compatible; MSIE 5.01; Windows NT 5.0)" You must use : LogFormat="%virtualname %host %other %logname %time1 %methodurl %code %bytesd %refererquot %uaquot"
62.161.78.73 - - [dd/mmm/yyyy:hh:mm:ss +0x00] "GET /page.html HTTP/1.1" 200 3904 "http://www.from.com/from.htm" "Mozilla/4.0 (compatible; MSIE 5.01; Windows NT 5.0)" mod_gzip: DECHUNK:OK In:11393 Out:3904:66pct. You must use : LogFormat="%host %other %logname %time1 %methodurl %code %bytesd %refererquot %uaquot %other %other %gzipin %gzipout"
62.161.78.73 - - [dd/mmm/yyyy:hh:mm:ss +0x00] "GET /page.html HTTP/1.1" 200 3904 "http://www.from.com/from.htm" "Mozilla/4.0 (compatible; MSIE 5.01; Windows NT 5.0)" (45) You must use : LogFormat="%host %other %logname %time1 %methodurl %code %bytesd %refererquot %uaquot" LogSeparator=" *"
200.135.30.181 - - [dd/mmm/yyyy:hh:mm:ss +0x00] "GET http://www.mydomain.com/page.html HTTP/1.0" 200 456 TCP_CLIENT_REFRESH_MISS:DIRECT You must use : LogFormat="%host %other %logname %time1 %methodurl %code %bytesd %other"
yyyy-mm-dd hh:mm:ss GET /page.html - 62.161.78.73 - Mozilla/4.0+(compatible;+MSIE+5.01;+Windows+NT+5.0) http://www.from.com/from.html 200 1234 HTTP/1.1 You must use : LogFormat="%time2 %method %url %logname %host %other %ua %referer %code %bytesd %other"
yyyy-mm-dd hh:mm:ss 62.161.78.73 - 192.168.1.1 80 GET /page.html - 200 11205 0 0 HTTP/1.1 mydomain.com Mozilla/4.0+(compatible;+MSIE+5.5;+Windows+98) - http://www.from.com/from.html You must use : LogFormat="%time2 %host %logname %other %other %method %url %other %code %bytesd %other %other %other %other %ua %other %referer"
62.161.78.73 - Name Surname Service [dd/mmm/yyyy:hh:mm:ss +0x00] "GET /page.html HTTP/1.1" 200 1234 "http://www.from.com/from.htm" "Mozilla/4.0 (compatible; MSIE 5.01; Windows NT 5.0)" You must use : LogFormat=6
62.161.78.73 - [dd/mmm/yyyy:hh:mm:ss +0x00] GET /page.html HTTP/1.1 200 1234 - "Mozilla/4.0 (compatible; MSIE 5.01; Windows NT 5.0)" Where separators are
See FAQ-COM100.
62.161.78.73 - - [dd/Month/yyyy:hh:mm:ss +0x00] "GET /page.html HTTP/1.1" "-" 200 1234 You must use : LogFormat="%host %other %logname %time1 %methodurl %other %code %bytesd" Note: Browsers, OS's, Keywords and Referers features are not available with a such format.
FAQ-SET270 : ONLY CORRUPTED OR DROPPED RECORDS PROBLEM: After running an AWStats update process, all my records are reported to be corrupted or dropped SOLUTION: First, if you have only a small percent of corrupted or dropped records, don't worry. This is a normal behaviour. Few corrupted or dropped records can appear in a log file because of internal web server bug, virus attack, error writing, log purge or rotate during a writing, etc... However, if ALL your records are reported to be corrupted or dropped, check the following things: If they are all dropped, run the update process from command line adding the option -showdropped -> You will be able to know why a dropped record is discarded. In most cases, this is because you use a too large or bad filter parameter (SkipFiles, SkipHosts, OnlyFiles ...). If they are all corrupted, run the update process from command line adding the option -showcorrupted -> You will be able to know why a corrupted record is discarded. If this is because of the log format, check the FAQ-SET350 about log format errors. If this is because the date of a record is said to be lower than date of previous, this means that you ran update processes on different log files without keeping the chronological order of log files. If this is because the date is invalid, you might have a problem of date not computed correctly this it happens in some Pentium4/Xeon4 processors: On some (few) Intel Pentium4 (also Xeon4) based host systems, log file time can not be computed correctly. This is not an issue of AWStats itself. This error usually occurs on source-based linux distributions (gentoo, slackware etc.), where all system libraries are compiled with CPU optimization. AWStats is a highly developed PERL application. PERL itself relies on some system libraries, for example GLIBC. The GLIBC library usually is buggy in this case. There is an easy way to figure out whether the problem described here is responsible for AWStats problems on your system: If you have shell access to your machine, simply type the following command: perl -e "print int ('541234567891011165415658')" (NOTE: any 25-digit number works, there is no need to type this exact number) If everything goes fine, you should see a floating point number as output: 5.41234567891011e+23 In this case, please do more research on your log file formats. Your host system itself is not responsible for the error. But if simply a "0" returns or some other error, this is an indication of your glibc beeing corrupt. ATTENTION: The only solution in this case might be to recompile your GLIBC. This can be a quite tricky task. Please consult the documentation and F.A.Q.s of your linux distribution first!! (experts: first check your global compile flags, eg. march=Pentium4. Trying with other compile flags can solve problem quickly in some cases. NOTE: In some cases, this error might occur "suddenly", even though AWStats was already running perfect already. FAQ-SET280 : ERROR "NOT SAME NUMBER OF RECORDS OF..." PROBLEM: When I run AWStats from command line (or as a cgi from a browser), I get a message "Not same number of records of ...". SOLUTION: This means your AWStats reference database files (operating systems, browsers, robots...) are not correct. First try to update to last version. Then check in your disk that you have only ONE of those files. They should be in 'lib' directory ('db' with 4.0) where awstats.pl is installed: browsers.pm domains.pm operating_systems.pm robots.pm search_engines.pm worms.pm status_http.pm status_smtp.pm FAQ-SET300 : ERROR "COULDN'T OPEN FILE ..." PROBLEM: I have the following error: "Couldn't open file /workingpath/awstatsmmyyyy.tmp.9999: Permission denied." SOLUTION: This error means that the web server didn't succeed in writing the working temporary file (file ended by .tmp.9999 where 9999 is a number) because of permissions problems. First check that the directory /workingpath has "Write" permission for user nobody (default user used by Apache on Linux systems) or user IUSR_SERVERNAME (default used user by IIS on NT). With Unix, try with a path with no links. With NT, you must check NTFS permissions ("Read/Write/Modify"), if your directory is on a NTFS partition. With IIS, there is also a "Write" permission attribute, defined in directory properties in your IIS setup, that you must check. With IIS, if a default cgi-bin directory was created during IIS install, try to put AWStats directly into this directory. If this still fails, you can change the DirData parameter to say AWStats that you want to use another directory (A directory you are sure that the default user, used by web server process, can write into). FAQ-SET320 : ERROR "MALFORMED UTF-8 CHARACTER (UNEXPECTED ..." PROBLEM: When running AWStats from command line, I get one or several lines like this on my output: Malformed UTF-8 character (unexpected non-continuation byte 0x6d, immediately after start byte 0xe4) at /www/cgi-bin/lib/xxx.pm line 999. SOLUTION: This problem appeared with RedHat 8 and Perl 5.8. I don't know if RedHat provides a fix for this, but some users had reported that you can remove thoose warmless messages by changing your LANG environment variable, removing the ".UTF-8" at the end. For example, set LANG="en_US" instead of LANG="en_US.UTF8" FAQ-SET350 : EMPTY OR NULL STATISTICS REPORTED PROBLEM: AWStats seems to work but I'm not getting any results. i get a statistics page that looks like i have no hits. SOLUTION: That's one of the most common problem you can get and there is 3 possible reasons : 1) Your log file format setup might be wrong. If you use Apache web server The best way of working is to use the "combined" log format (See the Setup and Use page to know the way to change your Apache server log from "common" log format into "combined"). Don't forget to stop Apache, reset your log file and restart Apache to make change into combined effective. Then you must setup your AWStats config file with value LogFormat=1. If you want to use another format, read the next FAQ to have examples of LogFile value according to log files format. If you use IIS server or Windows built-in web server The Internet Information Server default W3C Extended Log Format will not work correctly with AWStats. To make it work correctly, start the IIS Snap-in, select the web site and look at it's Properties. Choose W3C Extended Log Format, then Properties, then the Tab Extended Properties and uncheck everything under Extended Properties. Once they are all unchecked, check off the list given in the Setup and Use page ("With IIS Server" chapter). You can also read the next FAQ to have examples of LogFormat value according to log files format. 2) You are viewing stats for a year or month when no hits was made on your server. When you run awstats, the reports is by default for the current month/year. If you want to see data for another month/year you must: Add -year=YYYY -month=MM on command line when building the html report page from command line. Use an URL like http://myserver/cgi-bin/awstats.pl?config=xxx&year=YYYY&month=MM if viewing stats with AWStats used as a CGI. 3) When you read your statistics, AWStats does not use the same config file than the one used for the update process. Scan your disk for files that match awstats.*conf and remove all files that are not the config file(s) you need (awstats.conf files, if found, can be deleted. It is better to use a config file called awstats.mydomain.conf). FAQ-SET360 : STATISTICS REPORTED EXCEPT FOR OS, BROWSERS, ROBOTS AND KEYWORDS/KEYPHRASES PROBLEM: AWStats seems to report my statistics however some charts, like robots, os', browsers, search engines, or keywords/keyphrases are empty. SOLUTION: If only robots, search engines or keywords/keyphrases are empty, this simply means your web site was not yet visited by any robots and noone found your site using a search engines (this happens particularly for Intranet which are not referenced on search engines). If all of them are empty or with only unknown values, even after several updates, this probably means that your logfile does not contains all informations, this happens with Apache when using the standard "common" log format instead of the standard "combined" log format. You may also use LogFormat=4 into your AWStats config files instead of 1. Read AWStats setup documentation to known how to setup your Apache Web server to report logs in a "combined" log format then set LogFormat=1 into your AWStats config file. FAQ-SET400 : PIPE REDIRECTION TO A FILE GIVE ME AN EMPTY FILE PROBLEM: I want to redirect awstats.pl output to a file with the following command : > awstats.pl -config=... [other_options] > myfile.html But myfile.html is empty (size is 0). If i remove the redirection, everythings works correctly. SOLUTION: This is not an AWStats bug but a problem between perl and Windows. You can easily solve this running the following command instead: > perl awstats.pl -config=... [other_options] > myfile.html FAQ-SET450 : NO PICTURES/GRAPHICS SHOWN PROBLEM: AWStats seems to work (all data and counters seem to be good) but I have no image shown. SOLUTION: With Apache web server, you might have troubles (no picture shown on stats page) if you use a directory called "icons" (because of Apache pre-defined "icons" alias directory), so use instead, for example, a directory called "icon" with no s at the end (Rename your directory physically and change the DirIcons parameter in config file to reflect this change). FAQ-SET700 : MY VISITS ARE DOUBLED FOR OLD MONTH I MIGRATED FROM 3.2 TO 5.X PROBLEM: After having migrated an old history file for a month, the number of visits for this month is doubled. So the number of "visits per visitor" is also doubled and "pages per visit" and "hits per visit" is divided by 2. All other data like "pages", "hits" and bandwith are correct. SOLUTION: This problem occurs when migrating history files from 3.2 to 5.x. To fix this you can use the following tip (warning, do this only after migrating from 3.2 to 5.x and if your visit value is doubled). The goal is to remove the line in history file that looks like this YYYYMM00 999 999 999 999 where YYYY and MM are year and month of config file and 999 are numerical values. So if your OS is Unix/Linux grep -vE '^[0-9]{6}00' oldhistoryfile > newhistoryfile mv newhistoryfile oldhistoryfile And then run the migrate process again on the file. If your OS is windows and got cygwin You must follow same instructions than if OS is Unix/Linux BUT you must do this from a cygwin 'sh' shell and not from the DOS prompt (because the ^ is not understanded by DOS). And then run the migrate process again on the file. In any other case (in fact works for every OS) You must remove manually the line YYYYMM00 999 999 999 999 (must find one and only one such line) and then run the migrate process again on the file. FAQ-SET750 : AWSTATS RUN OUT OF MEMORY DURING UPDATE PROCESS WITH CYGWIN PERL PROBLEM: When I run the update process on a large log file with cygwin Perl, AWStats run out of memory but I am sure that I have enough memory to run AWStats according to the 'memory' column in benchmark chart available in AWStats documentation (benchmark page). SOLUTION: It might be a limit inside Cygwin Perl. Try to increase the Cygwin parameter heap_chunk_in_mb. FAQ-SET800 : AWSTATS SPEED/TIMEOUT PROBLEMS ? PROBLEM: When I analyze large log files, processing times are very important (Example: update process from a browser returns a timeout/internal error after a long wait). Is there a setup or things to do to avoid this and increase speed ? SOLUTION: You really need to understand how a log analyzer works to have good speed. There is also major setup changes you can do to decrease your processing time. See important advices in benchmark page. FAQ-SEC100 : CAN AWSTATS BE USED TO MAKE CROSS SITE SCRIPTING ATTACKS ? PROBLEM: If a bad user use a browser to make a hit on an URL that include a < SCRIPT > ... < /SCRIPT > section in its parameter, when AWStats will show the links on the report page, does the script will be executed ? SOLUTION: No. AWStats use a filter to remove all scripts codes that was included in an URL to make a Cross Site Scripting Attack using a log analyzer report page. FAQ-SEC150 : HOW CAN I PREVENT SOME USERS TO SEE STATISTICS OF OTHER USERS ? PROBLEM: I don't want a user xxx (having a site www.xxx.com) to see statistics of user yyy (having a site www.yyy.com). How can i setup AWStats for this ? SOLUTION: Take a look at the security page. FAQ-SEC200 : HOW TO MANAGE LOG FILES (AND STATISTICS) CORRUPTED BY 'WORMS' ATTACKS ? PROBLEM: My site is attacked by some worms viruses (like Nimba, Code Red...). This make my log file corrupted and full of 404 errors. So my statistics are also full of 404 errors. This make AWStats slower and my history files very large. Can I do something to avoid this ? SOLUTION: Yes. 'Worms' attacks are infected browsers, robots or server changed into web client that make hits on your site using a very long unknown URL like this one: /default.ida?XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX%40%50...%40%50 URL is generated by the infected robot and the purpose is to exploit a vulnerability of the web server (In most cases, only IIS is vulnerable). With such attacks, you will will always find a 'common string' in those URLs. For example, with Code Red worm, there is always default.ida in the URL string. Some other worms send URLs with cmd.exe in it. With 6.0 version and higher, you can set the LevelForWormsDetection parameter to "2" and ShowWormsStats to "HBL" in config file to enable the worm filtering nd reporting. However, this feature reduce seriously AWStats speed and the worms database (lib/worms.pm file) can't contain all worms signatures. So if you still have rubish hits, you can modify the worms.pm file yourself or edit your config file to add in the SkipFiles parameter some values to discard the not required records, using a regex syntax like example : SkipFiles="REGEX[^\/default\.ida] REGEX[\/winnt\/system32\/cmd\.exe]" |
|