Re: [MRBS-general] Summary so far: Slow save problem with MRBS

[Campbell M. <cam...@gm...>]

The six identical POSTs are normal.   I suspect they are the Ajax requests
to edit_entry_handler which get fired off by edit_entry every 10 seconds (or
when a field changes) to check that there are no schedule or policy
conflicts.    The return should result in two little green ticks (or red
crosses) next to the Save button on the edit_entry page.    Are you getting
these?    If not, then I suspect this has something to do with the problem
you are seeing.   If you're not getting these could you turn on Firebug and
see what the response is that you're getting?   Instead of getting a JSON
object you'll probably get some PHP error message in the response which will
give us a clue.

 

By the way, are you serving MRBS using https?    There is a bug in 1.4.8,
fixed in the trunk, that means that Report and Search don't work properly
when using https, but it should not affect edit_entry.  But it would be
interesting to know all the same.

 

My gut feeling at this stage is that it's probably not to do with the
database, but something to do with a JavaScript problem or else possibly a
rewrite rule in .htaccess.     I also suspect that you're probably not
getting as far as edit_entry_handler until the one and a half minutes is up.
I think the best thing to do is to put some debug lines such as echo "I am
here";  in various spots in  edit_entry_handler to find out where the delay
is.  I suspect that if you put such a line at the beginning of
edit_entry_handler.php it won't be reached for a minute or two.

 

 

Campbell

 

 

 

 

From: Craig Knights [mailto:cra...@gm...] 
Sent: 28 February 2012 22:52
To: DunLUG; General purpose list (support/developers/users)
Subject: [MRBS-general] Summary so far: Slow save problem with MRBS

 

Saving a new entry takes 1 to 2 minutes

 

Watching the apache log tells me that there are 6 identical POSTs evenly
spaced in that delay.. followed by the GET of the new data which is fast.
Is this normal?

 

>> 

web:80 10.161.51.96 - - [29/Feb/2012:11:35:30 +1300] "POST
/mrbs/edit_entry_handler.php HTTP/1.1" 302 493
"http://10.161.48.11/mrbs/edit_entry.php?area=5
<http://10.161.48.11/mrbs/edit_entry.php?area=5&room=31&period=0&year=2012&m
onth=2&day=28> &room=31&period=0&year=2012&month=2&day=28" "Mozilla/5.0
(Windows NT 6.1) AppleWebKit/535.11 (KHTML, like Gecko) Chrome/17.0.963.56
Safari/535.11"

>> 

 

 

I've changed all the tables to innoDB, no better.  (Jim and John)

 

I've got rid of the hostname warning I was getting on apache2 restarts.
Happens by default if you haven't set it.  

 

Accessing the page by the IP rather than the url doesn't help. (Me)

 

Email support is explicitly disabled in the config file for MRBS (Campbell)

 

The server is not busy (the system load is generally around 0.01), the pages
show up quickly in my browser, the network interface is idling.. (iftop,
top)

 

I didn't touch the server in the week before it starting acting up.

 

Dummy data INSERTs using phpmyadmin into the mrbs_entry table happen very
fast.

 

Slow on all 3 browsers tested (Chrome, Firefox and IE)

 

thanks,

Craig