Page 1 of 2 12 LastLast
Results 1 to 25 of 27
  1. #1
    Join Date
    Jul 2001
    Location
    Somewhere you're not
    Posts
    239

    Question Would Apache trouble affect ping times?

    Would overloaded Apache service slow down the ping times to a server?

  2. #2
    Its not going to be "apache" that causes high ping times. It would be either high load on the system, or saturated network. Apache its self wont reply to pings.

  3. #3
    Join Date
    Jul 2001
    Location
    Somewhere you're not
    Posts
    239
    OK, that's what I thought. But what kind of load are you talking about? A high amount of bandwidth?

  4. #4
    Load as in processor utilization. So if apache was hogging the processor, then yes, it could delay pings.

    Chris

  5. #5
    Join Date
    Jul 2001
    Location
    Somewhere you're not
    Posts
    239
    Oh, well yes it is hogging the processor very badly. Sometimes up to 99.9% of CPU resources. How can I determine if it's a configuration problem/bug with Apache that can be fixed, or if I need to upgrade my server's hardware?

  6. #6
    What operating systen are you running?

    What are your system specs?

    How my clients are on the server?


    Chris

  7. #7
    Join Date
    Jul 2001
    Location
    Somewhere you're not
    Posts
    239
    RedHat Linux 8.0
    WHM 9.4.0 cPanel 9.4.0-R21

    Dual Intel Xeon 2.4 Ghz
    1 GB DDR RAM
    73 GB SCSI drive

    The server is running one PHP/MySQL intensive site, with a large vBulletin forum.

  8. #8
    Thats a fairly hefty server...

    How much traffic does it push on a monthly basis?

    Do you have your kernel configured for multiple processors? I've seen that overlooked before.

    Can you past the output of top (type 'top' at ssh), that way we can see whats all running and see what state its in.

    Chris

  9. #9
    Join Date
    Jul 2001
    Location
    Somewhere you're not
    Posts
    239
    The server pushes about 300-500 GB per month.

    I don't know about the kernel, or how to check it...it's a server from a NOCSTER/BurstNET reseller. The server has ran great, with about the same amount of traffic, in the past.

    Code:
      2:27am  up 2 days,  9:04,  1 user,  load average: 1.68, 1.94, 2.21
    216 processes: 215 sleeping, 1 running, 0 zombie, 0 stopped
    CPU0 states:  5.0% user,  1.1% system,  5.2% nice, 93.0% idle
    CPU1 states:  7.0% user,  0.2% system,  7.3% nice, 91.4% idle
    CPU2 states: 13.2% user,  3.4% system, 13.2% nice, 82.1% idle
    CPU3 states:  6.0% user,  0.4% system,  6.3% nice, 92.2% idle
    Mem:  1030768K av, 1017472K used,   13296K free,       0K shrd,   94712K buff
    Swap: 1052216K av,  119304K used,  932912K free                  518120K cached
    
      PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM   TIME COMMAND
    24965 root      10   4  2676 1152  1116 S N  33.6  0.1  59:01 httpd
     2772 root      14   0  1240 1240   884 R     4.0  0.1   0:01 top
    14602 mysql     13   6 79352  62M 12404 S N   1.9  6.1  39:02 mysqld
        1 root       8   0   528  484   464 S     0.0  0.0   0:05 init
        2 root       9   0     0    0     0 SW    0.0  0.0   0:00 keventd
        3 root      19  19     0    0     0 SWN   0.0  0.0   0:00 ksoftirqd_CPU0
        4 root      19  19     0    0     0 SWN   0.0  0.0   0:00 ksoftirqd_CPU1
        5 root      19  19     0    0     0 SWN   0.0  0.0   0:00 ksoftirqd_CPU2
        6 root      19  19     0    0     0 SWN   0.0  0.0   0:00 ksoftirqd_CPU3
        7 root       9   0     0    0     0 SW    0.0  0.0   1:33 kswapd
        8 root       9   0     0    0     0 SW    0.0  0.0   0:00 bdflush
        9 root       9   0     0    0     0 SW    0.0  0.0   0:46 kupdated
       10 root       9   0     0    0     0 SW    0.0  0.0   0:00 xfsbufd
       11 root       9   0     0    0     0 SW    0.0  0.0   0:00 xfslogd/0
       12 root       9   0     0    0     0 SW    0.0  0.0   0:00 xfslogd/1
       13 root       9   0     0    0     0 SW    0.0  0.0   0:00 xfslogd/2
       14 root       9   0     0    0     0 SW    0.0  0.0   0:00 xfslogd/3
       15 root       9   0     0    0     0 SW    0.0  0.0   0:00 xfsdatad/0
       16 root       9   0     0    0     0 SW    0.0  0.0   0:00 xfsdatad/1
       17 root       9   0     0    0     0 SW    0.0  0.0   0:00 xfsdatad/2
       18 root       9   0     0    0     0 SW    0.0  0.0   0:00 xfsdatad/3

  10. #10
    Join Date
    May 2003
    Location
    Kirkland, WA
    Posts
    4,448
    Another gig of RAM definitely wouldn't hurt. You're swapping a bit.
    Nick Nelson
    Sr. Director & GM, VAS
    Demand Media
    425.298.2282 nn@demandmedia.com

  11. #11
    You are getting into swap space fairly hevily. Reboot your server to recover from this. Then have anouther 512 or 1024 megs of RAM added to the server.

    Also, Apache is running as root, which is a bad idea. You may want to consult a security specialist as well.

    Chris

  12. #12
    Join Date
    Jul 2001
    Location
    Somewhere you're not
    Posts
    239
    OK, thanks for the help, guys. If it was an Apache issue I was going to hire a professional admin to check on it for me, but if it can't be helped without more hardware then I guess that's just what I'll have to do.

  13. #13
    Join Date
    Jul 2001
    Location
    Somewhere you're not
    Posts
    239
    Just an update...I restarted my server, as you suggested, Chris, and my server immediately returned with a high load. There's not even swapping...or does it matter? Would I still need more RAM just to keep the load down?

    Code:
      2:48am  up 5 min,  1 user,  load average: 1.97, 1.34, 0.58
    140 processes: 139 sleeping, 1 running, 0 zombie, 0 stopped
    CPU0 states:  2.1% user,  8.1% system,  0.0% nice, 89.1% idle
    CPU1 states:  3.2% user,  0.3% system,  0.0% nice, 95.4% idle
    CPU2 states:  5.4% user,  0.5% system,  0.0% nice, 93.0% idle
    CPU3 states:  1.0% user,  0.4% system,  0.0% nice, 98.0% idle
    Mem:  1030768K av,  368152K used,  662616K free,       0K shrd,   36952K buff
    Swap: 1052216K av,       0K used, 1052216K free                  138584K cached
    
      PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM   TIME COMMAND
     1336 root       9   0  4928 4928  4760 S    14.1  0.4   0:14 httpd
     1116 mysql      9   0  9820 9820  1416 S     1.7  0.9   0:02 mysqld
     1619 root      11   0  1124 1124   884 R     0.3  0.1   0:00 top
      179 root       9   0     0    0     0 SW    0.1  0.0   0:00 kjournald
      182 root       9   0     0    0     0 SW    0.1  0.0   0:01 kjournald
      910 root       9   0  1384 1384  1248 S     0.1  0.1   0:00 sshd
        1 root       8   0   544  544   480 S     0.0  0.0   0:03 init
        2 root       9   0     0    0     0 SW    0.0  0.0   0:00 keventd
        3 root      19  19     0    0     0 SWN   0.0  0.0   0:01 ksoftirqd_CPU0
        4 root      18  19     0    0     0 SWN   0.0  0.0   0:00 ksoftirqd_CPU1
        5 root      19  19     0    0     0 SWN   0.0  0.0   0:00 ksoftirqd_CPU2
        6 root      18  19     0    0     0 SWN   0.0  0.0   0:00 ksoftirqd_CPU3
        7 root       9   0     0    0     0 SW    0.0  0.0   0:00 kswapd
        8 root       9   0     0    0     0 SW    0.0  0.0   0:00 bdflush
        9 root       9   0     0    0     0 SW    0.0  0.0   0:01 kupdated
       10 root       9   0     0    0     0 SW    0.0  0.0   0:00 xfsbufd
       11 root       9   0     0    0     0 SW    0.0  0.0   0:00 xfslogd/0
       12 root       9   0     0    0     0 SW    0.0  0.0   0:00 xfslogd/1
       13 root       9   0     0    0     0 SW    0.0  0.0   0:00 xfslogd/2
       14 root       9   0     0    0     0 SW    0.0  0.0   0:00 xfslogd/3
       15 root       9   0     0    0     0 SW    0.0  0.0   0:00 xfsdatad/0
       16 root       9   0     0    0     0 SW    0.0  0.0   0:00 xfsdatad/1
       17 root       9   0     0    0     0 SW    0.0  0.0   0:00 xfsdatad/2
       18 root       9   0     0    0     0 SW    0.0  0.0   0:00 xfsdatad/3
       20 root       9   0     0    0     0 SW    0.0  0.0   0:00 ahc_dv_0
       21 root       9   0     0    0     0 SW    0.0  0.0   0:00 scsi_eh_0
       23 root      -1 -20     0    0     0 SW<   0.0  0.0   0:00 mdrecoveryd

  14. #14
    Well, your processors look idle, so I'm guessing the load may be caused by the disk drives because you're certanly not using up your processors.

    Try something like:

    /sbin/hdparm -Tt /dev/hda

    That will perform a performance test on your disk drives. Might be able to eliminate or pinpoint a disk drive issue.

    Are you're databases well kept up and optimized?

    Chris

  15. #15
    Join Date
    Jul 2001
    Location
    Somewhere you're not
    Posts
    239
    OK, I tried running that command but it returned this...

    root@server [/]# /sbin/hdparm -Tt /dev/hda
    /dev/hda: No such device or address

    Yes, the databases and MySQL were just optimized by a vBulletin technician.

  16. #16
    I think you need to call somebody who has more or less decent knowledge in our business. Apparently your application on the top of HTTP misbehaved and this is not "nice".

    Peter Kinev
    Open Solution, Inc
    http://opensolution-us.com

  17. #17
    Join Date
    Jul 2001
    Location
    Somewhere you're not
    Posts
    239
    Is there anybody you would recommend, pnorilsk? You see, I'm not really in "your business" and I'm not a host. I just run my website as my hobby and I'm trying to run my own server as best I can.

  18. #18
    Originally posted by Lancia
    Is there anybody you would recommend, pnorilsk? You see, I'm not really in "your business" and I'm not a host. I just run my website as my hobby and I'm trying to run my own server as best I can.
    Fair enough. If you compare two outputs from the "top" command (needless to say it's instantaneous value in time and could be very much different the next moment, so some care in reading of output from the "top" must be in placed), you will see that "nice" (sorry) values for httpd and mysql on the first output are different from the second. The "nice" value is a relatively static entity, but could be changed to get a better priority (see output) at runtime. Apparently it was changed - for the system with httpd / mysql only, it's an idiotic idea to change it - the complete nonsense. So, with higher priority httpd and mysql taked awy all CPU cycles from any other process, including the response on your ping.

    Peter KInev.
    Open Solution, Inc
    http://opensolution-us.com

  19. #19
    One other thing that comes to mind, you are on Burst?

    Place like a 10meg file on the server and download it from anouther network and see what type of transfer rates you get. It could be your network is saturated, which would cause the higher ping times.

    As for the server loads, as pnorilsk suggested, you may want to have someone take a look at it. If SQL is hammering your disk drive, that could be causing the server load.

    Out of the two 'top' ourputs, I see two instances of loads around 2. Idealy, you should always try and keep your load under 3. Based on the fact that I've only seen 2 'top' outputs, its probably safe to say at peak times your server may go over a load of 2.

    You may want to occasionaly take a peak and see how well its doing, and if you see loads increasing, have someone take a look at your system.

    Chris

  20. #20
    Originally posted by EMT-Chris

    As for the server loads, as pnorilsk suggested, you may want to have someone take a look at it. If SQL is hammering your disk drive, that could be causing the server load.

    No, no, no. I didn't say anything about load. It's a special and different conversation. You are correct that "uptime" output what "top" is displaying could be every time different. The reason is, the "load average" displayed values for the last 1, 5, and 15 minutes.

    I was talking about "nice" and "priority" - different entities to compare with "load average".

    Peter Kinev.
    Open Solution, Inc
    http://opensolution-us.com

  21. #21
    Originally posted by pnorilsk
    No, no, no. I didn't say anything about load. It's a special and different conversation. You are correct that "uptime" output what "top" is displaying could be every time different. The reason is, the "load average" displayed values for the last 1, 5, and 15 minutes.

    I was talking about "nice" and "priority" - different entities to compare with "load average".

    Peter Kinev.

    I understand. I only stated that I agreed that he should have someone take a look at the server. There is only so much we can do from a forum. Someone should get a birds eye view of his system configuration. The rest of that paragraph was to simply inform the OP he should keep an eye on the load to see what it does at different points during the day.

    Sorry for any confusion.

    Chris

  22. #22
    Originally posted by EMT-Chris
    I understand. I only stated that I agreed that he should have someone take a look at the server. There is only so much we can do from a forum. Someone should get a birds eye view of his system configuration. The rest of that paragraph was to simply inform the OP he should keep an eye on the load to see what it does at different points during the day.

    Sorry for any confusion.

    Chris
    Chris,

    Agree, that what I suggested few postings back. Still, can you speculate what or who "reniced" priority? I know Apache (did some work), server will not "renice" it for sure. This is why I speculated about some thread/process of httpd child.

    Peter.
    Open Solution, Inc
    http://opensolution-us.com

  23. #23
    Originally posted by pnorilsk
    Chris,

    Agree, that what I suggested few postings back. Still, can you speculate what or who "reniced" priority? I know Apache (did some work), server will not "renice" it for sure. This is why I speculated about some thread/process of httpd child.

    Peter.
    Hmm...

    I wouldnt consider myself an expert in the arena, as I don't monkey with it too often. If memory serves, you can set priorities anywhere from -20 (being high) to 19 (being low). Not sure what apache's default is (possibly 10 or 9? as it may appear).

    To raise the the priority, I think the command is something like:

    nice $priority httpd
    nice -15 httpd (for example)

    Then to prevent the service from stopping,

    renice $priority -p $processid (I think renice uses processes id)
    renice -15 -p 2047 (for example)

    *just some info for the OP


    Its quite possible you're correct though. My only suggestion would be to raise its priority and test it out. As I've said, my knowledge in this area if fairly limited. I would need to brush up on the intricates to be of more use beyond this...

    Chris
    Last edited by EMT-Chris; 06-14-2004 at 04:03 PM.

  24. #24
    Chris,

    I am sorry. You misunderstood me. I know how this part works very well. I don't know what application on the top of HTTPD will change nice value. I would not be surprised if CPanel or anything like that doing that.

    Peter Kinev
    Open Solution, Inc
    http://opensolution-us.com

  25. #25
    I too am mistified at why they would fluctuate that much.

    I assumed he manually configured his services, as he is only running a single site. Personally I wouldnt bother with cpanel in this case.

    The priority of a service should remain static. Perhaps the OP should manually apply a priority to httpd and mysql and see if it sticks?

    Between the funky priorities, server loads, high pings which I guess he never did say what the pings were (more hten likley just a saturated network though), it may not be a bad idea to have an admin take a peak around.

    Chris

Page 1 of 2 12 LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •