Why this server so slow

Posted on 1997-08-14
Last Modified: 2013-12-23
Now We have a solaris 2.5.1 server which run sendmail,pop3,
http,informix database. but since it worked, it runs sometime slow sometime quick, the web is slow till time out sometime, so does sendmail,pop3. I really don't why and how to deal with it.
Question by:gaofei
Welcome to Experts Exchange

Add your voice to the tech community where 5M+ people just like you are talking about what matters.

  • Help others & share knowledge
  • Earn cash & points
  • Learn & ask questions

Expert Comment

ID: 1582270

   Well .. it's really hard to solve your problem without
knowing the hardware you're running on, the memory. First,
you seems to be running too much stuff on it ... If the
average load on your server is more than 5 ...then I think you
will need to split the load on another server. When it's slow .. try to check how many processes are running ... same for the it's fast. Check which one is taking the most CPU & memory.
   The command "ps" will show you the process, "w" will show you
the current users & load, "netstat" will show you current network
    Performance tunning is a large issue ... of which I cannot
go into details. The commands above will help you understand
which process is taking a huge chunk of your CPU time, and what
make your network slow.
Good luck,
Minh Lai


Author Comment

ID: 1582271
I have use 'vmstat' to see the cpu and memory resource,but I
think system resource is enough,I also use 'ps' to see the
processes,but the processes are alse not more than that system
can deal with.
because all users are database users not unix users,so there
not so much users.So I don't know why this server so slow much
more,Please give me another answer.

Expert Comment

ID: 1582272
I'm just guessing here, but you say the users of the database are not unix users, so I guess you are using ODBC to access the system.  In my own experience I've seen that once you get more or less two thirds of the users to just log in to the database it slows the whole system down.

What happens is that informix sees the connection and accepts is once the user enters the password (in the case of odbc this is done automatically) and then claims a percentage of the unix kernel resources just in case that user decides to do some work.  What I would suggest is to try something: Get all the users to log out (or do this a night)...bounce the machine and see what happens, what the performance looks like.  And then again get ALL (or most) of the users to just log in to the database, then look at the memory and CPU usage (possibly swap file as well).  If I correct you may see a huge decrease in kernel resources.  Should this be the case I suggest going back to your supplier of Informix and see whether there are any patches available to solve this.

Free NetCrunch network monitor licenses!

Only on Experts-Exchange: Sign-up for a free-trial and we'll send you your permanent license!

Here is what you get: 30 Nodes | Unlimited Sensors | No Time Restrictions | Absolutely FREE!

Act now. This offer ends July 14, 2017.


Expert Comment

ID: 1582273
I too run an Informix DB among the many apps on Sol2.5.1 for a large company with a huge load with without any performance  problems. The easiest way to find out stats is pull up a performance meter and set it to log and follow it for any trends. The other thing is if the actual system is fine is that it could be the network. Howe many people log into this and how much are they pulling and through what kind of network device (TBase10, etc). The other thing is that it might be as stupid as that the database needs to be set up for performance. Do you use raw partitions, RAID (and if so what kinds?), or a standard UFS? What is the size of the database? It may need some archiving, compression, or purging... And you might want to take a look at swaping for the DB itself. Another thing is to take a look for a variable (which I can't remember off the top of my head) that is something like MAXTCP. You might be running out the maximum number of network transactions at a time the system can handle per cycle. Take a look at Adrian Cockroft's book on Solaris System Performance Tuning or O' Reilly's book on System Performance Tuning.

Accepted Solution

MalicEInd earned 200 total points
ID: 1582274
From experience running an ODBC database on a Solaris 2.5.1 Sparcstation-5 (64MB RAM), I also had a similar problem prior to installing all recommended patched (up to 1/7/97 at the time the problem arose).  I found informix was subject to the Solaris kernel mutex_enter bug, which caused significant memory leakage (when the initial authentication connection to the database was made)
You can probably test if this is the same situation by hitting the Informix database with rapid authentication requests.

Whilst doing this, watch the system load at frequent intervals (with sar - probably 2 seconds would show a good indication of load).  The processor load should hit 100% reasonably quickly, and stay there even after authentication is complete.  Also monitor swap/memory access, you should find that memory usage jumps rapidly, then appears to drop down quite rapidly.  (Though not all of the memory threads are correctly released).


Expert Comment

ID: 1582275
its so annoying when people dont grade questions :)


Expert Comment

ID: 1582276
its so annoying when people dont grade answers :(


Author Comment

ID: 1582277
but I really don't understang you all, how can I repair this
mutex_enter bug?


Author Comment

ID: 1582278
but I really don't understand you all, how can I repair this
mutex_enter bug?


Featured Post

Technology Partners: We Want Your Opinion!

We value your feedback.

Take our survey and automatically be enter to win anyone of the following:
Yeti Cooler, Amazon eGift Card, and Movie eGift Card!

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

If you're not part of the solution, you're part of the problem.   Tips on how to secure IoT devices, even the dumbest ones, so they can't be used as part of a DDoS botnet.  Use PRTG Network Monitor as one of the building blocks, to detect unusual…
In this article, I am going to show you how to simulate a multi-site Lab environment on a single Hyper-V host. I use this method successfully in my own lab to simulate three fully routed global AD Sites on a Windows 10 Hyper-V host.
Here's a very brief overview of the methods PRTG Network Monitor ( offers for monitoring bandwidth, to help you decide which methods you´d like to investigate in more detail.  The methods are covered in more detail in o…
In this tutorial you'll learn about bandwidth monitoring with flows and packet sniffing with our network monitoring solution PRTG Network Monitor ( If you're interested in additional methods for monitoring bandwidt…
Suggested Courses

691 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question