Announcement

Collapse
No announcement yet.

mod_lsapi errno 104

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • mod_lsapi errno 104

    Hello there,

    I have a issue with Apache mod_lsapi.
    Apache is crashing non stop on my server and my

    > /usr/local/apache/logs/error_log

    file is filling up with this error message:

    Code:
    [Wed Mar 08 22:55:01.509429 2017] [lsapi:error] [pid 446729:tid 139975853004544] [client XX.XX.XX.XX:59059] [host XXX.COM] Backend error on sending request(GET /support/XXXXXX.php HTTP/1.0); uri(/index.php) content-length(0) (lsphp is killed?): ReceiveAckHdr: backend process reset connection: errno 104 (possibly lsapi_backend_children parameter too small)
    
    fork() failed, please increase process limit: Cannot allocate memory
    
    [Wed Mar 08 22:43:04.827956 2017] [lsapi:error] [pid 446610:tid 139975911753472] [client XX.XX.XX.XX:51614] [host [URL]http://www.XXX-XXX.de[/URL]] Backend error on sending request(GET /images/product_images/info_images/628_0.jpg HTTP/1.0); uri(/index.php) content-length(0) (lsphp is killed?): ReceiveAckHdr: nothing to read from backend socket, referer: [URL]http://www.XXXX.XXXX/xxx-xxxxx/[/URL]
    
    Out of memory
    As in docs under Troubleshouting explained I set lsapi_backend_children to 10000 but it changed nothing and Apache keeps crashing.

    As I can guess something is wrong with memory, but I dont have a clue where I should start to debug.

    At all the server have enough memory available.

    Code:
    free -g
    
    total        used        free      shared  buff/cache   available
    
    Mem:            125          11           8           6         105         105
    
    Swap:            15           0          15
    Apache is running under around 8 to 10 requests/sec.

    This is my first setup of cPanel + CL7 with mod_lsapi + optimumcache so I dont have any experience how much load this setup can handle. My other server are all running under dso handler with more than double of the load without issues like this.

  • #2
    There are two possible reasons for such issue.

    1. You have set RLimitMem somewhere in apache config, please check it out with:

    Code:
    grep RLimitMem /usr/local/apache/conf* -Rs
    2. Your LVE memory limits are to low for normal website functionality, check if any LVEs are hitting limits with:

    Code:
    lveinfo --by-fault --period=12h

    Comment


    • #3
      Thank you Bogdan for the fast replay.

      I checked for RLimitMem but there is no setting like this. I also disabled the Apache Memory Usage Restrictions function in cPanel beforr.

      lveinfo isnt showing something special neither:

      Code:
      lveinfo --by-fault mem --period=12h
      
      +----+----+----+----+---+---+---+------+-----+-----+-----+---+------+------+-----+------+------+------+-----+------+----+-----+------+-----+-----+-----+
      
      | ID |aCPU|mCPU|lCPU|aEP|mEP|lEP|aVMem |mVMem|lVMem|VMemF|EPf|aPMem |mPMem |lPMem|aNproc|mNproc|lNproc|PMemF|NprocF|aIO | mIO | lIO  |aIOPS|mIOPS|lIOPS|
      
      +----+----+----+----+---+---+---+------+-----+-----+-----+---+------+------+-----+------+------+------+-----+------+----+-----+------+-----+-----+-----+
      
      |1058|0   |3   |100 |0  |0  |100|1.2MB |256MB|256MB|1    |0  |278KB |17.3MB|256MB|0     |1     |100   |0    |0     |99B |8.8KB|10.0MB|0    |1    |1.0K |
      
      |1526|0   |12  |100 |0  |0  |100|22.8MB|256MB|256MB|16   |0  |8.1MB |34.0MB|256MB|0     |1     |100   |0    |0     |523B|140KB|10.0MB|0    |1    |1.0K |
      
      |1582|0   |6   |100 |0  |0  |100|9.7MB |256MB|256MB|11   |0  |3.6MB |31.4MB|256MB|0     |1     |100   |0    |0     |93B |845B |10.0MB|0    |0    |1.0K |
      
      |1584|0   |6   |100 |0  |1  |100|15.6MB|256MB|256MB|547  |0  |5.6MB |28.3MB|256MB|0     |1     |100   |0    |0     |140B|1.3KB|10.0MB|0    |0    |1.0K |
      
      |1588|0   |7   |100 |0  |0  |100|20.1MB|256MB|256MB|11   |0  |7.5MB |31.3MB|256MB|0     |1     |100   |0    |0     |182B|1.3KB|10.0MB|0    |0    |1.0K |
      
      |1610|1   |9   |100 |0  |1  |100|27.1MB|256MB|256MB|4.3K |0  |10.3MB|32.4MB|256MB|0     |2     |100   |0    |0     |855B|6.8KB|10.0MB|0    |0    |1.0K |
      
      |1627|1   |9   |100 |0  |0  |100|23.5MB|256MB|256MB|18   |0  |8.8MB |31.5MB|256MB|0     |1     |100   |0    |0     |253B|2.8KB|10.0MB|0    |0    |1.0K |
      
      |1643|0   |5   |100 |0  |0  |100|20.1MB|256MB|256MB|147  |0  |7.6MB |28.4MB|256MB|0     |1     |100   |0    |0     |229B|1.8KB|10.0MB|0    |0    |1.0K |
      
      |1664|0   |7   |100 |0  |2  |100|10.5MB|256MB|256MB|363  |0  |4.0MB |39.9MB|256MB|0     |3     |100   |0    |0     |147B|1.1KB|10.0MB|0    |0    |1.0K |
      
      |1666|0   |1   |100 |0  |0  |100|10.1MB|256MB|256MB|66   |0  |3.8MB |28.3MB|256MB|0     |1     |100   |0    |0     |85B |961B |10.0MB|0    |0    |1.0K |
      
      +----+----+----+----+---+---+---+------+-----+-----+-----+---+------+------+-----+------+------+------+-----+------+----+-----+------+-----+-----+-----+

      Comment


      • #4
        The lveinfo does show needed information. There were a lot of vmem faults (VMemF column) for those accounts. Setting it to 256 Mb is to low.

        Even more, we strongly recommend disabling vmem limits completely setting it to 0 , you have pmem limit enabled so you are good to go.

        Comment


        • #5
          Indeed this issue was produced by vmem.
          You should put this advice with big bold letters into the documentation as Apache is really weird and unpredictable.

          Thank you very much for your help!

          Comment

          Working...
          X