I am running openldap v 1.2.10 on a linux redhat 6.0, and I really experience a lot of problems... Indeed, the ldapserver crashes with the same message calloc of 57672698 elems of 4 bytes failed (i launch slapd with the -1debug level) This occurs when I try to insert such person objects : cn=vignaux eric, ou=DR AMIENS, o=sacem, c=fr I don t know what to do now, because i tried to upgrade my version of ldap, compile it without thread support... The only thing i can do is re generating the database with an ldif file, at this point, it works ok, but one later, problems occur Any idea ? thank s and apologize for my english, eric vignaux
At 05:16 PM 5/19/00 GMT, eric.vignaux@sacem.fr wrote: >calloc of 57672698 elems of 4 bytes failed Make sure you have enough virtual memory available to your process to allocate configured caches. Since calloc() is failing, allocate more virtual memory to your process and/or reduce cache sizes. See cachesize and dbcachesize slapd.conf(5) directives.
My swap partition had been fixed at 128 MB at the installation, and i have just reduced cachesize (1000 entries) and dbcachesize (100000 b) in my slapd.conf. But problems continue to occur. Could it be a ldbm backend problem ? Do you have any idea ? > -----Message d'origine----- > De: Kurt@OpenLDAP.org [SMTP:Kurt@OpenLDAP.org] > Date: dimanche 21 mai 2000 16:54 > À: openldap-its@OpenLDAP.org > Objet: Re: ldapadd problems...slapd crashes ... (ITS#540) > > At 05:16 PM 5/19/00 GMT, eric.vignaux@sacem.fr wrote: > >calloc of 57672698 elems of 4 bytes failed > > Make sure you have enough virtual memory available > to your process to allocate configured caches. Since > calloc() is failing, allocate more virtual memory > to your process and/or reduce cache sizes. See > cachesize and dbcachesize slapd.conf(5) directives. >
changed notes changed state Open to Feedback
At 07:56 AM 5/23/00 GMT, eric.vignaux@sacem.fr wrote: >But problems continue to occur. You are likely still exceeding the virtual memory available to the process. You should allow the process to allocate more virtual memory (which could be a matter of changing soft or hard limits or system resources) or further decrease cache settings. You might also have to reduce the number of indices maintained. >Could it be a ldbm backend problem ? It could be a number of things. However, the problem is likely due to resource limitations. >Do you have any idea ? % limit (check process limits)
changed state Feedback to Closed
resource limitation?