Optimizing Plesk

[fastcooler]

Hallo,

I bought a VPS running Virtuozzo and Plesk 7.5.4 Reloaded. I would like to know if is there a way to optimize Plesk. In fact system resources are always over 85%...

I have only 2 domains, but actually they are on a private area, so they aren't accessible from visitors.

Any ideas, please?

Thanks
Fast

 

[jamesyeeoc]

If your server is running a RedHat distro, it is normal for the resources/utilization (such as RAM) to be high. This is not a Plesk issue, it is the normal for RH systems.

Example: I have a test server which sitting at idle (with 1GB RAM) shows 90%+ RAM utilization under RedHat OS

Even my RH servers without Plesk do the same. (like my plain DNS boxes, no hosting or other things, only function is as a nameserver)

 

[fastcooler]

Hi James,

yes OS is Fedora, and these are system's informations:

  Primary UBC Parameters 
ID Current Use Soft Limit Hard Limit Units Description   
numproc 41 n/a 240 pcs Number of processes and kernel-level threads allowed for VPS  
numtcpsock 20 n/a 500 pcs Number of TCP sockets  
numothersock 19 n/a 360 pcs Number of non-TCP sockets (local sockets, UDP and other types of sockets)  
vmguarpages 0 32,768 2,147,483,647 4KB pages Memory allocation guarantee  
avnumproc n/a n/a 40 pcs Average estimated number of processes  
 
  Secondary UBC Parameters 
ID Current Use Soft Limit Hard Limit Units Description   
kmemsize 2,423,785 9,871,360 10,158,080 bytes Size of unswappable memory, allocated by the operating system kernel  
tcpsndbuf 0 1,800,000 3,400,000 bytes Total size of buffers used to send data over TCP network connections  
tcprcvbuf 0 1,800,000 3,400,000 bytes Total size of buffers used to temporary store the data coming from TCP network connections  
othersockbuf 23,136 1,126,080 2,097,152 bytes Total size of buffers used by local connections and send buffers of UDP and other datagram protocols  
dgramrcvbuf 0 262,144 262,144 bytes Total size of buffers used to temporary store the incoming packets of UDP and other datagram protocols  
oomguarpages 26,300 26,112 2,147,483,647 4KB pages The guaranteed amount of memory for the case the memory is "over-booked" (out-of-memory kill guarantee)

Is there a way to disable SPAMASSASSIN from loading at startup, please?

I saw that if I manually stop it from shell, system's resources go down at 48% that it's a very good result for me.

But in if I stop it manually I can't get HORDE Webmail to work... so I think that they are connected each other...

May you help me, please?
Thanks
Regards
Fast

 

[Cranky]

You can *remove* spamassassin from your VPS via:

rpm -e psa-spamassassin spamassassin

But you probably don't need to do that. Run the following via shell:

cat /proc/user_beancounters

And paste the output here (put it inside PHP tags so I can read it clearly when posting please, i.e.:

here

 

[fastcooler]

Hi Cranky!

this is output requested:

Version: 2.5
       uid  resource           held    maxheld    barrier      limit    failcnt
      1066: kmemsize        2224171    2227708    9871360   10158080          0
            lockedpages           0          0        256        256          0
            privvmpages       43558      43558      65536      65536          0
            shmpages           9123       9123      21846      21846          0
            dummy                 0          0          0          0          0
            numproc              40         40        240        240          0
            physpages         21052      21054          0 2147483647          0
            vmguarpages           0          0      32768 2147483647          0
            oomguarpages      21052      21054      26112 2147483647          0
            numtcpsock           19         19        500        500          0
            numflock              4          4        188        206          0
            numpty                1          1         16         16          0
            numsiginfo            2          2        256        256          0
            tcpsndbuf          4440       4440    1800000    3400000          0
            tcprcvbuf             0        428    1800000    3400000          0
            othersockbuf      16904      16904    1126080    2097152          0
            dgramrcvbuf           0          0     262144     262144          0
            numothersock         16         16        360        360          0
            dcachesize       176681     179909    2273280    2416640          0
            numfile             561        562       5820       5820          0
            dummy                 0          0          0          0          0
            dummy                 0          0          0          0          0
            dummy                 0          0          0          0          0
            numiptent            10         10        128        128          0

I hope this may help!
Thanks!
Best Regards
Fast

 

[fastcooler]

Hello,

nobody may help me, please?
Thanks
Regards
Fast

 

[Cranky]

You don't have much memory, perhaps ask your provider for a price on an upgrade? Alternatively, you could remove spamassassin if you really wanted to. You may be better placed posting this in the Virtuozzo forum too.

 

[MaRiOs]

how can I see the results of user_beancounters in another OS ? (like RHEL3) ?

I tried to do locate user_beancounters and didnt find anything..
and its not in the proc/ path :-(

 

[jamesyeeoc]

Marios - That is the command to see it on RH type servers.

You are not finding it in /proc probably since you are not running on a VPS type server.

 

[Cranky]

Marios, you would only see it on Virtuozzo-based VPS's. I was suffering from lack of sleep so I've nicely confused everyone in this thread.