do not upgrade to mm-1.1.3-8

Discussion in 'PSA 5.0 Troubleshooting and Problems' started by Linulex, Aug 19, 2002.

  1. Linulex

    Linulex Product Expert

    Messages:
    723
    do not upgrade to mm-1.1.3-8 on RH7.1

    The list up upgrade rpm's has mm-1.1.3-8 in it. You don't want to upgrade this version of mm. Stick with 1.1.3-2

    Here is the reason why

    http://forum.plesk.com/showthread.php?s=&threadid=4533

    It took me 2 sleepless nights to figure this one out.

    Kind regards

    Jan
  2. shorty

    shorty Guest

    I would be interested if anyone else has had a problem with 1.1.3-8 as we have it running on quite a few servers without problem and the old mm does have a local exploit which could be dangerous http://lwn.net/Alerts/6370/
  3. Dan S.

    Dan S. Guest

    The problems described in that thread must have been due to something else. I have succesfully installed mm-1.1.3-8 on a production RH 7.1 server, and stop/started apache with no problems.
  4. moefid

    moefid Guest

    plesk rpm 2.5.3 rh71, 72 & 73

    mm-1.1.3-8

    no problem
  5. Linulex

    Linulex Product Expert

    Messages:
    723
    The server starts and stops without a problem. Its just after restarting it about 50 times it wont restart again with that error.

    And even is somebody only changes his/hers ftp password then apache is restarted. So on a server where a lot of people activly use the interface it willbe sonner then on a server with few accounts where people never use the interface.

    regards

    Jan
  6. shorty

    shorty Guest

    ok we duplicated your problem - we added ftp users and deleted them on our test server, I think just over 60 times :) and then we got the error as in your other post.

    Just looking into the cause now and will update if we can find a solution other than downgrading
  7. shorty

    shorty Guest

    Linulex what does this give you on your system?

    cat /proc/sys/kernel/shmmax
  8. Linulex

    Linulex Product Expert

    Messages:
    723
    -
    I get

    33554432

    system is redhat 7.1 with all updates and kernel 2.4.9-34 standard rh kernel from the update. One server still has 2.2.4-2 and has the same problem after upgrading. with mm 1.1.3-2 there is no problem.

    ulimts is:

    > ulimit -a
    core file size (blocks) 0
    data seg size (kbytes) unlimited
    file size (blocks) unlimited
    max locked memory (kbytes) unlimited
    max memory size (kbytes) unlimited
    open files 1024
    pipe size (512 bytes) 8
    stack size (kbytes) 8192
    cpu time (seconds) unlimited
    max user processes 16384
    virtual memory (kbytes) unlimited

    I have tried raising the openfiles but that didn't help.


    regards

    Jan
  9. shorty

    shorty Guest

    looks like our systems are identical - I have raised shmmax to 67108864 using this command:

    echo 67108864 >/proc/sys/kernel/shmmax

    I'm still re-running the test on creating deleting ftp accounts (seems the quickest way to restart httpd in a semi-real way) and so far no failure, I will keep you informed to see if this has had any eefect
  10. frostie

    frostie New Member

    Messages:
    74
    So seeing as MM causes problems is there anyway to revert to an older version? And if so what are the steps that need to be taken to do so.
    Thanks in advance for your help.

    James
  11. shorty

    shorty Guest

    doesn't seem to be any of the old MM rpm's around, but the old MM has an exploit too!
  12. frostie

    frostie New Member

    Messages:
    74
    Always encouraging :p

    James
  13. Linulex

    Linulex Product Expert

    Messages:
    723
    mm-1.1.3-2 rpm can be found on the rh 7.1 cd.

    An exploit is usualy more a name then a reality. in 99% of the exploits you need to be a pretty good programmer to be able to really use them. i don't mean copy/paste the exploit script and then see if it works, but actually do some harm with it.

    Offcourse, if the upgrade/fix works fine then you should always upgrade.

    regards

    Jan
  14. shorty

    shorty Guest

    yeah we are getting this on production servers now - 3 today

    Starting httpd: Ouch! ap_mm_create(1048576, "/var/run/httpd.mm.2571") failed
    Error: MM: mm:core: failed to acquire semaphore (No space left on device): OS: Invalid argument
  15. shorty

    shorty Guest

    I've checked all my copies of red hat 7.1 and can't find mm on it, if you have a copy Linulex I would be grateful to get my hands on it :)
  16. Linulex

    Linulex Product Expert

    Messages:
    723
    send your email to jan@linulex.com and i'll mail them to you. They are about 40 K together.

    regards

    Jan
  17. shorty

    shorty Guest

    thanks Jan - lets hope that fixes it :)
  18. Linulex

    Linulex Product Expert

    Messages:
    723
    it did for me.

    don't forget to restart apache after; I also did restart the server just to be sure and clean the memory.

    regards

    Jan
  19. DarkHorizons

    DarkHorizons New Member

    Messages:
    257
    Can you tell me if these fixes worked for you?

    I am getting these stupid semaphore errors even after upgrading to 5.0.2 and they are driving me NUTS.

    If you have this issued nailed down can you please post the steps to fix it?

    THANKS!
  20. Linulex

    Linulex Product Expert

    Messages:
    723
    downgrading to mm 1.1.3-2 worked for me, had no problem with 2.5 anymore and still no problem with 5.0.2

    regards

    Jan

Share This Page