As part of the service, you can access your site's raw logfiles. The access logs are in Apache combined log format. If you wish to analyse your own logfiles, you can download these.
The CGI logs are error logs of script outputs. If you're running a script and it doesn't work, the standard error output will be written to these logfiles.
There are several machines in our server cluster writing to this platform, so you may need to allow a few seconds for today's files to update -- particularly if you are trying to debug a script.
- How do I access my logs?
- What logs are available?
- Do the logs affect my site quota?
- What do the logfiles mean?
- What programs can I analyse my logs with?
How do I access my logs?
You have two choices as to how you can access your logfiles -- either via a browser or any FTP program.
If you wish to collect them via a browser, you first need to log into our Self Care interface and then select the logfiles option from what you are presented on screen. When you have accessed your logfiles pages, you can then download them to your local machine. DO NOT bookmark this page - it uses the same mechanism as Webalizer stats - you can only view them if you have logged into Self Care first.
Here is how the Selfcare interface looks:
And you just click on the hyperlinks to view your logs.
Here is what a sample logfile looks like:
If you log into your webspace via an FTP program, you will see a directory called "/logs/" in your home directory. This has a subdirectory of "/logs/today/". When you login via FTP, you are actually put into the "/htdocs" directory by default. You need to go up one directory, and then you will see the "/logs/" directory.
What logs are available?
Today's logs in "/logs/today/" are the live access logs and CGI error logs (the ones that are being written to now by the web servers). The error logs contain just the stderr (Standard Error) output for your site's scripts. Here is where you look if you need information to debug your scripts.
In "/logs/today/", there is a log per backend web server - these are merged nightly into ../logs and compressed.
The files in "/logs/" contain the zipped access and CGI logs for the last 0 days for your site. These are rotated by the Webalizer program -- old logs are deleted daily.
GNU Zipped files can be uncompressed by most standard compression utilities, such as WinZip which you can get from Tucows.
Do the logs affect my site quota?
No. The logs are "owned" by a different user than yourself. Quotaing on the system is done by username, rather than by location of files (i.e. in which directory files are stored). This is also why you cannot delete or rename your logfiles.
What do the logfiles mean?
The access logs are in Combined Log Format. There is an explanation as to how this format is formed, and what the individual (space separated) fields mean available on the Apache website (even though we are using the Zeus web server and not the Apache one, the files are of the same format).
The CGI logs are the "standard error" of what your scripts output. Basically this means that if your script breaks, this contains the error message that is reported, along with other useful data such as the time that the error occured. This helps you bug fix your scripts (scripting coming in the next phase of the PWP project).
What programs can I analyse my logs with?
There are lots of programs on the web that will analyse standard Apache format logfiles. Amongst these are:
- Analog and its graphical interface Report Magic
- The Webalizer
These are both free to use. There are also programs you can buy that do similar jobs, but these are the most popular.
