Ridiculous vB Thread View Counts?

[Joe 90]

Can anyone help me with a fix... yesterday we had some new threads started that have like 5 replies or something... probably a little more than that actual views but they're displaying as like 3,500 views!?!

I've looked at the 'update counters' stuff and that doesn't seem to have fixed anything.

any ideas?

[Burble]

Sure it's not a bot crawling and bumping up the viewcount? If you've got access to the webserver logs grep out the thread ID and see roughly how many times it occurs.

[Daz]

Pipe it through wc -l :)

If it's happening a lot it'd be worth disabling plugins to see if that shuts it up. Any file hacks?

[Joe 90]

Pipe it through wc -l

:confused:


Its in a hidden forum, only about 7 users have access and the only thing I've got that might be able to access it is google analytics I think.
although I got a robots file from google which basically allows all;
User-agent: *
Allow: /

can robots crawl through authenticated forum areas though?

and what logs should I be looking at? the vB logs or actual server logs?

[Daz]

Robots will only see what guests see.

Look through the server logs. If you have shell access to the server, running

grep <threadid> access.log
will print out all the lines which match the thread id. Piping the result through wc, like this:

grep <threadid> access.log |wc -l
Will tell you how many lines (ie, how many requests) there were.

[Joe 90]

cheers daz.

ran a search across all log files including the error logs and found 427 lines.

Is there some way of archiving these results to download? my command line knowledge spreads about as far as navigating directories and using the help file!

[Daz]

grep <threadid> access.log > threadaccess.log
Replace threadaccess.log with whatever filename you like, and that will create a file containing each matching line for you to takeaway :)

[Joe 90]

awesome, thanks.

I really need to learn how to use this stuff properly!

what should i be lookin for in these logs btw!!? :o

[Garp]

Depends on how the server is set up to log. Typical format would be something like

IP, date, request, result code, size, source page, browser details.

try using head logfile.log to see the first 10 lines of it and check they're in that format. Then you could do something like:
cat logfile.log | awk {'print $1'} | sort -n |uniq -c | sort -n | tail
to see the ip address that's the biggest source of traffic. That command chain in plain english is: pass contents of log file to awk program, which returns first column ($1), which we then sort, and then produce a unique list with a count of occurrence, then finally sort them one last time, numerically, to put them in ascending count order, and return the final 10 lines.

From there you could do
grep "ip.ad.dre.ss" logfile.log | less tp grab a look at those entries.
The final column, the browser details, will reveal details like if it's Google spider that's crawling your site or similar. I would anticipate if it was it would appear a fair number of times in your logs.

[Joe 90]

well the heres a random line from the search of the logs;
ex090903.log:forum.shuwarriors.net: [03/Sep/2009:22:59:45 +0100] 91.107.220.54 - - "GET /clear.gif HTTP/1.1" 200 43 "http://forum.shuwarriors.net/showthread.php?t=1784" "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-GB; rv:1.9.1.2) Gecko/20090729 Firefox/3.5.2"

From what I said re:what i'm looking for though, how do I find something in all this that might explain the ridiculous number of views that the thread is returning?