Hi Daniel,

Ya at present Private memory is 1MB.
Shared memory is 1GB.
My Server is 32B.

btw Can you please let me know your suggestion how much private memory do I set?
As when I have set Private memory 4MB I was getting out of memory issue within 4 days.


Thanks,
Krunal Patel

On Fri, Jan 9, 2009 at 3:28 PM, Daniel-Constantin Mierla <miconda@gmail.com> wrote:
The dump shows a lot of free memory, about half of it:
Jan  7 22:25:36 ip-IP /usr/local/sbin/openser[23175]: Memory status (pkg):
Jan  7 22:25:36 ip-IP /usr/local/sbin/openser[23175]: qm_status (0x816cc00):
Jan  7 22:25:36 ip-IP /usr/local/sbin/openser[23175]:  heap size= 1048576  Jan  7 22:25:36 ip-IP /usr/local/sbin/openser[23175]:  used= 212192, used+overhead=545092, free=503484
Jan  7 22:25:36 ip-IP /usr/local/sbin/openser[23175]:  max used (+overhead)= 548088

However, the log is with openser having 1MB of private memory. Are you sure that you increased the memory pool to 4MB?

What was the size of shared memory? Is your server 32b or 64b?

Cheers,
Daniel



On 01/08/2009 11:44 AM, Krunal Patel wrote:
Hi,

I have captured latest memory dump using kill -SIGUSR1 OPENSER_PID.

Here is the links.

Link1 -- http://pastebin.com/m296598e2

Link2 -- http://pastebin.com/m9963b7c

Link3 -- http://pastebin.com/m7438e72a


As well as I have stopped openser & captured memory dump for the same.

Please find attached shutdown.tar.gz.


Please suggest me if there is any resolution.


Thanks in advance,


Krunal Patel

On Wed, Jan 7, 2009 at 2:31 PM, Daniel-Constantin Mierla <miconda@gmail.com <mailto:miconda@gmail.com>> wrote:

   Hello,


   On 01/07/09 07:26, Krunal Patel wrote:

       Hi,

       Was the signal SIGHUP or SIGUSR1?

       *++ The signal was SIGHUP1.*

   SIGHUP is 1, so i am confused whether there was a mistake in
   typing as there should be SIGUSR1.


       A cause can be insufficient private memory, as it seems you
       have a quite big config file, with lot of variables. In this
       case, the private memory used by config is relevant,
       considering that private memory per process is 1MB, for
       runtime is not that much left. Try to recompile openser with
       2MB of private memory and see if the situation occurs again.

       *++ I had increased private memory to 4MB eventhough the
       situation occurred after 4 days of restarting openser.
       *

       *++ Have you find any memory leak in the memory dump?*

   No, that log does not show a leak - there are memory chunks
   allocated at startup not at runtime. Please do again the tests and
   be sure you follow:
   - send SIGUSR1 to the process that log memory error messages, not
   to the main process or other processes
   - grab that log and send it, then you can stop openser (you will
   get another set of memory dump logs)

   Thanks,
   Daniel


       Thanks for you support,

       --
       Krunal Patel


       On Wed, Jan 7, 2009 at 4:35 AM, Daniel-Constantin Mierla
       <miconda@gmail.com <mailto:miconda@gmail.com>
       <mailto:miconda@gmail.com <mailto:miconda@gmail.com>>> wrote:

          Hello,


          On 01/01/09 10:42, Krunal Patel wrote:

              Hi,

              Here is the memory dump captured using kill -SIGHUP1
       OPENSERPID.
              Please help me out to interpret it.

              http://pastebin.com/m39a6c204
              http://pastebin.com/m518f27b7

          looks like all memory chunks were allocated at start up time.

          Was the signal SIGHUP or SIGUSR1?

          A cause can be insufficient private memory, as it seems you
       have a
          quite big config file, with lot of variables. In this case, the
          private memory used by config is relevant, considering that
          private memory per process is 1MB, for runtime is not that much
          left. Try to recompile openser with 2MB of private memory
       and see
          if the situation occurs again.

          Cheers,
          Daniel


              Thanks
              Krunal Patel



              On Wed, Dec 31, 2008 at 6:26 PM, Daniel-Constantin Mierla
              <miconda@gmail.com <mailto:miconda@gmail.com>
       <mailto:miconda@gmail.com <mailto:miconda@gmail.com>>
              <mailto:miconda@gmail.com <mailto:miconda@gmail.com>
       <mailto:miconda@gmail.com <mailto:miconda@gmail.com>>>> wrote:

                 Hello,

                 there is something wrong in your side the DBG_QM_MALLOG
              does not
                 show up in the flags. Here is mine:

                 # openser -V

                 version: openser 1.2.3-notls (i386/linux)
                 flags: STATS: Off, USE_IPV6, USE_TCP, DISABLE_NAGLE,
       USE_MCAST,
                 SHM_MEM, SHM_MMAP, PKG_MALLOC, DBG_QM_MALLOC,
              FAST_LOCK-ADAPTIVE_WAIT

                 ADAPTIVE_WAIT_LOOPS=1024, MAX_RECV_BUFFER_SIZE 262144,
              MAX_LISTEN
                 16, MAX_URI_SIZE 1024, BUF_SIZE 65535
                 poll method support: poll, epoll_lt, epoll_et, sigio_rt,
              select.
                 svnrevision: 2:5379M

                 After PKG_MALLOC, I have DBG_QM_MALLOC. Double check
       your
                 Makefile.defs.

                 Daniel



                 On 12/31/08 14:50, Krunal Patel wrote:

                     Hi

                     Here is the output:
                     version: openser 1.2.3-notls (i386/linux)
                     flags: STATS: Off, USE_IPV6, USE_TCP, DISABLE_NAGLE,
                     USE_MCAST, SHM_MEM, SHM_MMAP, PKG_MALLOC,
              FAST_LOCK-ADAPTIVE_WAIT
                     ADAPTIVE_WAIT_LOOPS=1024, MAX_RECV_BUFFER_SIZE
       262144,
                     MAX_LISTEN 16, MAX_URI_SIZE 1024, BUF_SIZE 65535
                     poll method support: poll, epoll_lt, epoll_et,
              sigio_rt, select.
                     svnrevision: unknown
                     @(#) $Id: main.c 3173 2007-11-20 08:26:35Z
       bogdan_iancu $
                     main.c compiled on 02:27:47 Dec 24 2008 with gcc
       4.0.2

                     Thanks,
                     Krunal Patel
         
--
Daniel-Constantin Mierla
http://www.asipto.com