Results 1 to 4 of 4
  1. #1
    Join Date
    Sep 2008
    Location
    Toronto,Canada
    Posts
    655

    Unanswered: How to start ids engine back ?

    In development env, ids was shut down with -ky switch ( onmode -ky ).

    While starting back with ( oninit ), it throws shared memory error :-

    08:12:44 shmget: [EEXIST][17]: key 52714801: shared memory already exists
    08:12:44 Could not create single shared memory segment with resident and non-resident par

    titions. Proceeding to create 2 shared memory segments instead.
    08:12:44 shmget: [EEXIST][17]: key 52714801: shared memory already exists
    08:12:44 mt_shm_init: can't create resident segment

    How to start it back ?? ids 7.31 on aix

    Thanks

  2. #2
    Join Date
    Sep 2011
    Location
    Pont l'Abbé, Brittany, France
    Posts
    381
    Provided Answers: 1
    Hi,

    maybe late, but you probably have zombie shared memory segments that prevent the engine from starting.

    Probably due to an unexpected crash ( you should check whether you have af.nnn.nnnn
    files before 08.:12:44

    In any case, what you have to do it to identify those shared memory segments by running

    ipcs -m
    identify segments whose keys look to be consecutive ( i-e beginning with the same first 6 character values),
    like this:

    Code:
    Key              shmid         owner      perms
    0x525d4801   24117275    root         660
    0x525d4802   24117275    root         660
    0x525d4803   24117275    root         660
    0x525d4804   24117275    root         600
    If you have only one instance in the server, you can shoot the first segment, that will probably shoot the consecutive ones, using ipcrm -m shmidvalue

    Caution, this operation can cause severe collateral damage if you don't shoot the right segments!

  3. #3
    Join Date
    Sep 2011
    Location
    Pont l'Abbé, Brittany, France
    Posts
    381
    Provided Answers: 1
    once those segments are shot, you can restart your server normally with oninit

  4. #4
    Join Date
    Sep 2008
    Location
    Toronto,Canada
    Posts
    655
    Quote Originally Posted by begooden-it View Post
    Probably due to an unexpected crash ( you should check whether you have af.nnn.nnnn
    files before 08.:12:44

    In any case, what you have to do it to identify those shared memory segments by running

    ipcs -m
    identify segments whose keys look to be consecutive ( i-e beginning with the same first 6 character values),
    like this:

    Code:
    Key              shmid         owner      perms
    0x525d4801   24117275    root         660
    0x525d4802   24117275    root         660
    0x525d4803   24117275    root         660
    0x525d4804   24117275    root         600
    If you have only one instance in the server, you can shoot the first segment, that will probably shoot the consecutive ones, using ipcrm -m shmidvalue

    Caution, this operation can cause severe collateral damage if you don't shoot the right segments!
    Thanks, for info.

    I had tried but could not identify segments; we have multiple engines running.

    I saw a few, but could not relate one to the crashed one, so I refrained from shooting out.

    Now the whole server has been bounced ( Saturday). So I lost the chance to try it.

    What had happened before was that the server was instable, still running, developers were unable to use it. Probably it had memory corruption; or whatever ?? It was shut down with -ky switch.

    Are these memroy segments always consecutive multiples or can be single consecutive or non-consecutive.
    What I mean is that : if 3 ids servers were started 5 mins apart, how many shared memory segments are allocated, and what kind of layout in the memory is expected ??

    Thanks anyway, info you provided makes sense and is abundant.

    DBFinder

Posting Permissions

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