stronghold dies with shmcb error

Hi,

  Recently I have noticed a problem in our stronghold web farm. Usually,
we have about 300 - 500 processes serving requests at this time of day;
when this problem ocurs, the number of httpd processes drops to around
10, and this node stops serving requests.

  A stop-server && start-server is useless, I have to kill all the httpd
processes by hand and start the server afterwards. Does anybody have a
clue on what this could be related (obviously some sort of limit in
mod_ssl session storage)?


[Thu Apr 21 16:32:24 2005] [alert] Child 22838 returned a Fatal error...
Apache is exiting!
[Thu Apr 21 16:32:24 2005] [error] mod_ssl: shmcb_insert_encoded_session,
internal error
[Thu Apr 21 16:32:24 2005] [error] mod_ssl: can't store a session!
[Thu Apr 21 16:32:24 2005] [error] mod_ssl: 'shmcb' code was unable to
store a
session in the cache.
[Thu Apr 21 16:32:24 2005] [error] mod_ssl: shmcb_insert_encoded_session,
internal error
[Thu Apr 21 16:32:24 2005] [error] mod_ssl: can't store a session!
[Thu Apr 21 16:32:24 2005] [error] mod_ssl: 'shmcb' code was unable to
store a
session in the cache.
[Thu Apr 21 16:32:34 2005] [error] mod_ssl: shmcb_insert_encoded_session,
internal error
[Thu Apr 21 16:32:34 2005] [error] mod_ssl: can't store a session!
[Thu Apr 21 16:32:34 2005] [error] mod_ssl: 'shmcb' code was unable to
store a
session in the cache.


Best of regards,
  Ricardo Oliveira