My newest site connected to mssql2 is unGAWDly slow this morning.
This is the exact same code that runs on my other sites (connected to mssql1) that appear to be functioning normally.
Something going on with mssql2 this morning?
Opened a ticket on it… EIX-13514-520
But don’t expect anything to be done until Stephen reads this post, as usual.
Site is now completing timing out. No just long load times, but SQL timeout errors.
Yay, fun!
And this is the response I get.
Seriously, this is not acceptable.
Simple ASP/HTML pages that aren’t connected to the SQL server load fine. Any page that relies on SQL for data IS UNGODLY SLOW!
The website was already set up with it’s own dedicated app pool, so that shouldn’t be the problem.
Again, only pages that connect to mssql2 are affected.
From: JodoHost Support [mailto:[email protected]]
Sent: Tuesday, March 10, 2009 8:35 AM
To: xxxxxxx
Subject: [JH #EIX-13514-520]: mssql2 issues/abuse
Hello,
The server seems to be working fine all the day. We dont find any such issue on the MSSQL server. Plesae update us with the domain name for which you are facing this problem.
Feel free to contact us for further assistance.
With Regards
Arijeet Neogi
there is actually not any abuse 4-15% on the server,it looks like osmeone was doing some large full text indexing before but it wasn’t really abusive.
I found part of the problem, looks like about 80 IPs trying to brute force the server making login troubles. I am working to block them.
Maybe you’d like to take a shot at explaining the SQL timeouts I was getting this morning then?
I’m not crazy Stephen, and I only open tickets / raise a stink here on the forums when something is broken and people aren’t willing to believe it.
This is another one of those instances. Something was going on this morning (and still isn’t back up to snuff), and the new site was in the crapper for a good 2 to 3 hours. It’s working better now than it was previously, but it’s still slow compared to sites connected to MSSQL1.
What would I need to do to have my database moved to MSSQl1? I’d much rather be on that server instead since I can still connect to that one with SQL Mgmt Studio to manage my data. I hate that new 3rd party SQL client piece of garbage I was told to use for MSSQL2.
MSSQL2 may have had timeouts due to the login attempts from many IPs as i mentioned.
There is no way to move MSSQL or MySQL DBs and maintain any control.
It’s happening again this morning.
Microsoft OLE DB Provider for SQL Server error ‘80004005’
Timeout expired
Compare load times between
and
The FaF site loads quickly. The Maelstrom site is garishly slow to respond, IF it responds at all.
Someone needs to put some effort into figuring this out and resolve it instead of just saying “looks fine to me”.
Just because it’s working “better” (IE, not as good as it should, but still better than when I’m reporting the problem) by the time your support staff gets around to looking at it DOES NOT mean it’s fixed or that there isn’t a problem.
Now I’m getting HTTP 500’s.
What the hell is going on with my siite?
So now my post has been “sent to the principals office” and moved out of public view?
This is getting mind numbingly ridiculous Stephen. This is the sort of thing that prompts people to look for other hosting providers.
SIX AND A HALF HOURS LATER and the site is still having issues.
What is going on?!?
@Phurion:
FYI, I just tried both sites. Both much the same to fully load:
furyandfire.net: 28 seconds
maelstromonline.com: 27 seconds
Neither are particularly fast. Refreshing either page takes between 5 and 9 seconds. I notice that the page generation time displayed at the bottom of the page varies greatly depending on how long since I last called the page. If I don’t refresh the page for 5 or more minutes, the page takes between 3.5 and 4.5 seconds to be generated. If I reload the page immediately, it takes 0.1 to 0.3 seconds to generate. The figures are similar across both sites,
I know a lot of CMSs use temporary caching to speed up page access. Could it be that furyandfire.net has more visits and therefore the chances of fresh data cached are higher than with maelstromonline.com?
I know nothing about SOOP Portal but it might be worth checking out if/how it caches pages.
The FaF site has been retired and has only 2 to 3 visitors per day. I moved the guild community over to the Maelstrom site, which is the problem site.
Initial load times on the FaF site were always long, until the web server could establish a connection with the SQL server, but subsequent page loads were quick, generally less than 3 seconds.
This morning the Maelstrom site is loading pretty fast.
This page was generated in 1.7637 seconds.
We’ll see how long that lasts.
The initial page load on the FaF site was 8 seconds, but subsequent page refreshes are showing normal render times.
This page was generated in 1.9375 seconds.
SOOP uses simple java image pre-loads for cacheing, nothing fancy.
Prior to this morning (and may yet still) I was getting SQL time outs on the Maelstrom site more than I was getting successful page loads. If a page DID complete loading it was usually several minutes in the process.