openldap-technical

openldap-technical@openldap.org

9 participants
9942 discussions

OpenLDAP log file ownership getting modified automatically.
by Prashanth P.Nair 01 Apr '16

01 Apr '16

Hello Am using OpenLDAP 'openldap-2.4.31' on debian. I have configured syncrepl .Everything work’s normal. But frequently am getting below Error. The log file ownership is getting changed any one of the node automatically . here is the Error From syslog. Mar 29 12:14:26 xxx slapd[29405]: conn=5774 op=40651 SEARCH RESULT tag=101 err=0 nentries=1 text= Mar 29 12:14:29 xxx slapd[29405]: bdb(dc=xxx,dc=xxx,dc=xxx): /var/lib/ldap/xxx/log.0000000883: log xxxle unreadable: Permission denied Mar 29 12:14:29 xxx slapd[29405]: bdb(dc=xxx,dc=xxx,dc=xxx): PANIC: Permission denied Mar 29 12:14:29 xxx slapd[29405]: bdb(dc=xxx,dc=xxx,dc=xxx): DB_ENV->log_newfh: 883: DB_RUNRECOVERY: Fatal error, run database recovery Mar 29 12:14:29 xxx slapd[29405]: bdb(dc=xxx,dc=xxx,dc=xxx): txn_checkpoint: log failed at LSN [883 68682]: DB_RUNRECOVERY: Fatal error, run database recovery Mar 29 12:14:30 xxx slapd[29405]: conn=5774 op=40652 SRCH base="" scope=0 deref=3 xxxlter="(objectClass=*)" Mar 29 12:14:30 xxx slapd[29405]: conn=5774 op=40652 SRCH attr=objectclass Any idea what is going wrong here? LDAP is running as below user . openldap:x:106:108:OpenLDAP Server Account,,,:/var/lib/ldap:/bin/false $ id openldap uid=106(openldap) gid=108(openldap) groups=108(openldap) I dont find any logs for open file issue.But still i pasted below settings. open file limit settings openldap soft nofile 65536 openldap hard nofile 65536 slapd.conf settings loglevel sync stats modulepath /usr/lib/ldap moduleload back_hdb moduleload syncprov sizelimit 500 tool-threads 8 threads 16 backend hdb database hdb monitoring on cachesize 50000 idlcachesize 50000 directory "/var/lib/ldap/account" dbconfig set_cachesize 0 167772160 0 dbconfig set_lk_max_objects 1500 dbconfig set_lk_max_locks 1500 dbconfig set_lk_max_lockers 1500 index objectClass,snAccount,snEnabled,entryCSN,entryUUID eq index cn eq,pres,subany lastmod on checkpoint 5120 30 access to dn.base="" by * read syncrepl rid=001 provider=ldap://xx.xx.xxx.xxx bindmethod=simple binddn="cn=xxx,ou=administrators,dc=test,dc=IN" credentials=xxx searchbase="dc=account,dc=test,dc=IN" schemachecking=on type=refreshAndPersist retry="10 +" mirrormode on overlay syncprov syncprov-checkpoint 100 10 syncprov-sessionlog 10000 Br/Prashanth

3 8

Re: Making a full replica: slapd -c "rid=xxx" doesn't seem to work
by Jephte Clain 01 Apr '16

01 Apr '16

hello, I haven't tried full (config + data) replication from a multimaster setup, but I think that doing so, the slave would become a master like the other from which it replicate, wouldn't it? cc-ing the list to have the opinion of the real experts :-) by the way, as I write this, I recall an advice from Michael Ströder: if the slaves are to replicate everything from the masters, why wouldn't you make them masters like the others? multi-master replication also work for more than two masters. best regards, Jephté Clain 2016-04-01 18:32 GMT+04:00 Jussi Sallinen <jussi(a)synack.fi>: > Ah, > > This is two multimaster servers so I guess that approach doesn't work, as it's meant for master/slave? > > -- > -Jussi Sallinen | UNIX System Administrator > jussi(a)synack.fi | Syn Ack > +358 40 700 7600 | Helsinki, Finland > >> On 1.4.2016, at 11.08, Jephte Clain <jephte.clain(a)univ-reunion.fr> wrote: >> >> hello, >> >> one more thing: make sure to have the proper syncprov overlay AND >> syncrepl directive pointing to the master slapd on the master, at >> least on olcDatabase={0}config and olcDatabase={1}mdb (or whatever is >> the first database that contains the data), e.g: >> >> dn: olcDatabase={0}config,cn=config >> olcSyncrepl: {0}rid=0 provider="ldap://master/" searchbase=cn=config ... >> olcUpdateRef: ldap://master/ >> ... >> >> dn: olcOverlay={0}syncprov,olcDatabase={0}config,cn=config >> ... >> >> dn: olcDatabase={1}mdb,cn=config >> olcSuffix: dc=univ-reunion,dc=fr >> olcSyncrepl: {0}rid=1 provider="ldap://master/" searchbase=dc=univ-reunion,dc=fr >> olcUpdateRef: ldap://master/ >> ... >> >> dn: olcOverlay={0}syncprov,olcDatabase={1}mdb,cn=config >> ... >> >> the syncrepl directive does nothing on the master, but as this >> configuration is replicated on the slave, the slave knows it has to >> replicate from the master (I hope it's clear enough) >> >> >> >> 2016-04-01 11:06 GMT+04:00 Jussi Sallinen <jussi(a)synack.fi>: >>> Hi, >>> >>> Thank you for the reply! >>> That's what I thought could be the answer. >>> >>> I will now need to try out how this works with our actual database dc=synack,dc=fi >>> >>> Basicly since this will be a new replica without even the dcDomain object in the initial database it should pick up all the data from existing multimaster replica (empty database is without csn etc and contains no update timestamp). >>> >>> -- >>> -Jussi Sallinen | UNIX System Administrator >>> jussi(a)synack.fi | Syn Ack >>> +358 40 700 7600 | Helsinki, Finland >>> >>>> On 1.4.2016, at 8.27, Jephte Clain <jephte.clain(a)univ-reunion.fr> wrote: >>>> >>>> Le 01/04/2016 01:26, Jussi Sallinen a écrit : >>>>> Hi, >>>>> >>>>> How you managed to get this working or did you come up with a workaround: http://www.openldap.org/lists/openldap-technical/201201/msg00205.html >>>>> >>>>> I have a similar situation where I'm doing a reprovision of my 2nd multimaster replica virtualhost and adding base config etc. with ldapadd (we build it from scratch with Ansible during reprovisioning). >>>>> After the new multimaster replica comes up it has newer timestamps than the current up & running replica from where this new one is supposed to replicate everything as we use ldapadd during the reprovision phase with Ansible. >>>> >>>> hello, >>>> >>>> after you create the seed replica, you just have to "touch" the master to ensure the timestamp on the master is newer than on the replica. >>>> >>>> As the seed configuration only contains 3 objects: cn=config, olcDatabase={-1}frontend,cn=config and olcDatabase={0}config,cn=config, I do it that way: >>>> >>>> $ ldapmodify -H ldap://master ...credentials... <<EOF >>>> dn: cn=config >>>> changetype: modify >>>> >>>> dn: olcDatabase={-1}frontend,cn=config >>>> changetype: modify >>>> >>>> dn: olcDatabase={0}config,cn=config >>>> changetype: modify >>>> EOF >>>> >>>> best regards, >>>> Jephté Clain >>>> >>>> -- >>>> *Jephté CLAIN | Développeur, Intégrateur d'applications* >>>> Service Systèmes d'Information >>>> Direction des Systèmes d'Information <http://dsi.univ-reunion.fr> >>>> Tél: +262 262 93 86 31 <tel:+262262938631> || Gsm: +262 692 29 58 24 <tel:+262692295824> >>>> www.univ-reunion.fr <http://www.univ-reunion.fr> || Facebook <http://www.facebook.com/pages/Universit%C3%A9-de-La-R%C3%A9union-OFFICIEL/1…> || Twitter <http://twitter.com/univ_reunion> >> >> >> >> -- >> Jephté CLAIN | Développeur, Intégrateur d'applications >> Service Système d'Information >> Direction des Systèmes d'Information >> Tél: +262 262 93 86 31 || Gsm: +262 692 29 58 24 >> > -- Jephté CLAIN | Développeur, Intégrateur d'applications Service Système d'Information Direction des Systèmes d'Information Tél: +262 262 93 86 31 || Gsm: +262 692 29 58 24

1 0

Need help unpicking stats logging
by Philip Colmer 01 Apr '16

01 Apr '16

I've currently got stats logging turned on while I try to troubleshoot an application and I've noticed some rather strange searches going on. Strange in that the searches are for very high uidNumber values or for uid values that don't exist ... suggesting that someone might be trying to grab data from our server. What I'm struggling with is trying to figure out from the logs (a) the IP address that these queries are coming from and/or (b) the authenticated account being used (even if anonymous). For example, if I have a log line like this: conn=1928683 op=24 SRCH base="ou=accounts,dc=linaro,dc=org" scope=2 deref=0 filter="(&(uid=tftp)(objectClass=posixAccount)(&(uidNumber=*)(!(uidNumber=0))))" is there anything I can do with the conn or op values to connect that particular search query to an earlier logged BIND log entry? Or is there a different/better way for me to try and get the information I'm after? Thanks. Philip

2 1

Re: Need help unpicking stats logging
by Philip Colmer 01 Apr '16

01 Apr '16

Many thanks, Emmanuel, for your quick reply. That is certainly helping me figure out what is going on. It looks like one or two of the client systems aren't correctly configured and this is helping me pin them down. Regards Philip On 1 April 2016 at 11:49, Emmanuel Lécharny <elecharny(a)gmail.com> wrote: > Le 01/04/16 12:45, Philip Colmer a écrit : >> I've currently got stats logging turned on while I try to troubleshoot >> an application and I've noticed some rather strange searches going on. >> Strange in that the searches are for very high uidNumber values or for >> uid values that don't exist ... suggesting that someone might be >> trying to grab data from our server. >> >> What I'm struggling with is trying to figure out from the logs (a) the >> IP address that these queries are coming from and/or (b) the >> authenticated account being used (even if anonymous). >> >> For example, if I have a log line like this: >> >> conn=1928683 op=24 SRCH base="ou=accounts,dc=linaro,dc=org" scope=2 >> deref=0 filter="(&(uid=tftp)(objectClass=posixAccount)(&(uidNumber=*)(!(uidNumber=0))))" >> >> is there anything I can do with the conn or op values to connect that >> particular search query to an earlier logged BIND log entry? > > Just grep your log for 'conn=1928683', it will give you all the history > for the 'user' that did the search. > > Now, FTR, that serach really looks like a PAM LDAP or a SSSD search. > Pretty standard. >

1 0

Re: Making a full replica: slapd -c "rid=xxx" doesn't seem to work
by Jephte Clain 01 Apr '16

01 Apr '16

hello, one more thing: make sure to have the proper syncprov overlay AND syncrepl directive pointing to the master slapd on the master, at least on olcDatabase={0}config and olcDatabase={1}mdb (or whatever is the first database that contains the data), e.g: dn: olcDatabase={0}config,cn=config olcSyncrepl: {0}rid=0 provider="ldap://master/" searchbase=cn=config ... olcUpdateRef: ldap://master/ ... dn: olcOverlay={0}syncprov,olcDatabase={0}config,cn=config ... dn: olcDatabase={1}mdb,cn=config olcSuffix: dc=univ-reunion,dc=fr olcSyncrepl: {0}rid=1 provider="ldap://master/" searchbase=dc=univ-reunion,dc=fr olcUpdateRef: ldap://master/ ... dn: olcOverlay={0}syncprov,olcDatabase={1}mdb,cn=config ... the syncrepl directive does nothing on the master, but as this configuration is replicated on the slave, the slave knows it has to replicate from the master (I hope it's clear enough) 2016-04-01 11:06 GMT+04:00 Jussi Sallinen <jussi(a)synack.fi>: > Hi, > > Thank you for the reply! > That's what I thought could be the answer. > > I will now need to try out how this works with our actual database dc=synack,dc=fi > > Basicly since this will be a new replica without even the dcDomain object in the initial database it should pick up all the data from existing multimaster replica (empty database is without csn etc and contains no update timestamp). > > -- > -Jussi Sallinen | UNIX System Administrator > jussi(a)synack.fi | Syn Ack > +358 40 700 7600 | Helsinki, Finland > >> On 1.4.2016, at 8.27, Jephte Clain <jephte.clain(a)univ-reunion.fr> wrote: >> >> Le 01/04/2016 01:26, Jussi Sallinen a écrit : >>> Hi, >>> >>> How you managed to get this working or did you come up with a workaround: http://www.openldap.org/lists/openldap-technical/201201/msg00205.html >>> >>> I have a similar situation where I'm doing a reprovision of my 2nd multimaster replica virtualhost and adding base config etc. with ldapadd (we build it from scratch with Ansible during reprovisioning). >>> After the new multimaster replica comes up it has newer timestamps than the current up & running replica from where this new one is supposed to replicate everything as we use ldapadd during the reprovision phase with Ansible. >> >> hello, >> >> after you create the seed replica, you just have to "touch" the master to ensure the timestamp on the master is newer than on the replica. >> >> As the seed configuration only contains 3 objects: cn=config, olcDatabase={-1}frontend,cn=config and olcDatabase={0}config,cn=config, I do it that way: >> >> $ ldapmodify -H ldap://master ...credentials... <<EOF >> dn: cn=config >> changetype: modify >> >> dn: olcDatabase={-1}frontend,cn=config >> changetype: modify >> >> dn: olcDatabase={0}config,cn=config >> changetype: modify >> EOF >> >> best regards, >> Jephté Clain >> >> -- >> *Jephté CLAIN | Développeur, Intégrateur d'applications* >> Service Systèmes d'Information >> Direction des Systèmes d'Information <http://dsi.univ-reunion.fr> >> Tél: +262 262 93 86 31 <tel:+262262938631> || Gsm: +262 692 29 58 24 <tel:+262692295824> >> www.univ-reunion.fr <http://www.univ-reunion.fr> || Facebook <http://www.facebook.com/pages/Universit%C3%A9-de-La-R%C3%A9union-OFFICIEL/1…> || Twitter <http://twitter.com/univ_reunion> > -- Jephté CLAIN | Développeur, Intégrateur d'applications Service Système d'Information Direction des Systèmes d'Information Tél: +262 262 93 86 31 || Gsm: +262 692 29 58 24

1 0

Re: Making a full replica: slapd -c "rid=xxx" doesn't seem to work
by Jephte Clain 31 Mar '16

31 Mar '16

Le 01/04/2016 01:26, Jussi Sallinen a écrit : > Hi, > > How you managed to get this working or did you come up with a > workaround: > http://www.openldap.org/lists/openldap-technical/201201/msg00205.html > > I have a similar situation where I'm doing a reprovision of my 2nd > multimaster replica virtualhost and adding base config etc. with > ldapadd (we build it from scratch with Ansible during reprovisioning). > After the new multimaster replica comes up it has newer timestamps > than the current up & running replica from where this new one is > supposed to replicate everything as we use ldapadd during the > reprovision phase with Ansible. hello, after you create the seed replica, you just have to "touch" the master to ensure the timestamp on the master is newer than on the replica. As the seed configuration only contains 3 objects: cn=config, olcDatabase={-1}frontend,cn=config and olcDatabase={0}config,cn=config, I do it that way: $ ldapmodify -H ldap://master ...credentials... <<EOF dn: cn=config changetype: modify dn: olcDatabase={-1}frontend,cn=config changetype: modify dn: olcDatabase={0}config,cn=config changetype: modify EOF best regards, Jephté Clain -- *Jephté CLAIN | Développeur, Intégrateur d'applications* Service Systèmes d'Information Direction des Systèmes d'Information <http://dsi.univ-reunion.fr> Tél: +262 262 93 86 31 <tel:+262262938631> || Gsm: +262 692 29 58 24 <tel:+262692295824> www.univ-reunion.fr <http://www.univ-reunion.fr> || Facebook <http://www.facebook.com/pages/Universit%C3%A9-de-La-R%C3%A9union-OFFICIEL/1…> || Twitter <http://twitter.com/univ_reunion>

1 0

Should I be able to select an 'orphaned' DN from a translucent overlap proxy?
by Sullivan, Daniel [AAA] 31 Mar '16

31 Mar '16

Hi, I am writing to confirm the expected behavior of a translucent overlay proxy. I have the proxy working and can filter from the local database using the olcTranslucentLocal configuration option (I can see a merged record). My question pertains to ‘orphaned data’. For example, I can arbitrarily add a record to the local database by DN using ldapadd, the glue records are created and I can see the record in the output of slapcat. My problem is that a search via ldapsearch does not return this record unless their is a matching DN in the remote database, even if I am filtering by an attribute specifed in olcTranslucentLocal . Is this the expected behavior? There is more that one reason why I want to do this, but the lowest common denominator is that people are going to be moving records around upstream and I’d like to keep the local database tidy by blowing away records that no longer have a matching remote DN. On that note, the second reason why I’m interested in doing this is to create an individual default group for each user in the local database. I was planning on creating an actual group for each user programmatically, and have no problem doing this. Is there a more elegant best practice way to facilitate this sort of thing (i.e. an overlay solution), or is just creating the groups the way to go? I am using openldap-ltb.x86_64 2.4.44-2.el6. I appreciate your time and expertise. Dan ******************************************************************************** This e-mail is intended only for the use of the individual or entity to which it is addressed and may contain information that is privileged and confidential. If the reader of this e-mail message is not the intended recipient, you are hereby notified that any dissemination, distribution or copying of this communication is prohibited. If you have received this e-mail in error, please notify the sender and destroy all copies of the transmittal. Thank you University of Chicago Medicine and Biological Sciences ********************************************************************************

1 0

Question about ldap_init()/ldap_initialize() behavior with OpenLDAP
by Frank Crow 31 Mar '16

31 Mar '16

I'm not clear on whether I'm allowed to ask "C" API questions here. If not, I apologize in advance and please disregard in that case. However, the question is more about configuration than coding. I'm seeing that the ldap_init() function allows the hostname parameter to be null or a list of hosts. If it is null, the function attempts to determine the default ldap host. How does it determine the default LDAP host when the hostname parameter is null? Does it load the list of hosts from the ldap.conf URI parameter? And/or does it use the LDAPURI environment variable? Thanks, -- Frank

4 7

CSN is matching but entries count differs in master and slaves
by scn_73＠yahoo.com 31 Mar '16

31 Mar '16

Hello all, CSN is matching with master and slaves but if i count number of entries it's different. I am replicating complete DIT so count should be equal in Master and slave. Should i diff the missed entries and catch them in replications logs. Please suggest .. $ ldapsearch -H ldaps://ldapmaster -D "cn=manager,dc=xxx,dc=com" -x -xxx -s base contextCSN -LLL dn: dc=xxx,dc=com contextCSN: 20160326180539Z#000006#00#000000 $ ldapsearch -H ldaps://ldap-slave -D "cn=manager,dc=xxx,dc=com" -x -wxx -s base contextCSN -LLL dn: dc=xxx,dc=com contextCSN: 20160326180539Z#000006#00#000000 $ ldapsearch -H ldaps://ldapmaster -D "cn=manager,dc=xxx,dc=com" -x -wxx -s sub "objectclass=*" | tail -2 # numResponses: 28751 # numEntries: 28750 $ ldapsearch -H ldaps://ldap-slave "cn=manager,dc=xxx,dc=com" -x -wxxx -s sub "objectclass=*" | tail -2 # numResponses: 28772 # numEntries: 28771

3 4

ACL problem
by Cole 29 Mar '16

29 Mar '16

Hi, I am using OLC and ACL's to limit response from LDAP based on IP address, and this is working fine if I have a single ACL. When I add a second ACL, with a different IP address, the new ACL works, but the previous one does not. Example: Entries in LDAP: # 10.0.0.92, servers, test.com dn: cn=10.0.0.92,ou=servers,dc=test,dc=com objectClass: groupOfNames cn: 10.0.0.92 description: Allowed access member: uid=aaa,dc=test,dc=com # 10.0.0.94, servers, test.com dn: cn=10.0.0.94,ou=servers,dc=test,dc=com objectClass: groupOfNames cn: 10.0.0.94 description: Allowed access member: uid=aaa,dc=test,dc=com # aaa, test.com dn: uid=aaa,dc=test,dc=com objectClass: inetOrgPerson objectClass: posixAccount objectClass: ldapPublicKey objectClass: shadowAccount uid: aaa uidNumber: 10005 gidNumber: 10005 homeDirectory: /home/aaa loginShell: /usr/local/bin/bash mail: a(a)test.com ou: users cn: User A sn: A userPassword:: xxx OLC entries: # {1}ldif, config dn: olcDatabase={1}ldif,cn=config objectClass: olcLdifConfig olcDatabase: {1}ldif olcDbDirectory: /var/pq/0/mount/resolve/installations/entities/vmi/location/za /parts/new/instances/test/openldap-data olcSuffix: dc=test,dc=com olcAccess: {0}to filter=(|(&(objectClass=inetOrgPerson)(memberOf=cn=10.0.0.92, ou=servers,dc=test,dc=com))(objectClass=posixGroup)) by self read by peername .ip="10.0.0.92" read by * none olcAccess: {1}to * by self write by peername.ip="127.0.0.1" write by * none olcRootDN: cn=Manager,dc=test,dc=com olcRootPW: test At this point, running ldapsearch from the computer with ip address 10.0.0.92 works correctly, and only the filtered results are returned. If I then add a second ACL: # {1}ldif, config dn: olcDatabase={1}ldif,cn=config objectClass: olcLdifConfig olcDatabase: {1}ldif olcDbDirectory: /var/pq/0/mount/resolve/installations/entities/vmi/location/za /parts/new/instances/test/openldap-data olcSuffix: dc=test,dc=com olcAccess: {0}to filter=(|(&(objectClass=inetOrgPerson)(memberOf=cn=10.0.0.94, ou=servers,dc=test,dc=com))(objectClass=posixGroup)) by self read by peername .ip="10.0.0.94" read by * none olcAccess: {1}to filter=(|(&(objectClass=inetOrgPerson)(memberOf=cn=10.0.0.92, ou=servers,dc=test,dc=com))(objectClass=posixGroup)) by self read by peername .ip="10.0.0.92" read by * none olcAccess: {2}to * by self write by peername.ip="127.0.0.1" write by * none olcRootDN: cn=Manager,dc=test,dc=com olcRootPW: test The computer with IP 10.0.0.94 can now query LDAP, and receives the correct filtered results, however the computer with IP 10.0.0.92 receives a blank result: $ ldapsearch -x -h 10.0.0.91 -b dc=test,dc=com # extended LDIF # # LDAPv3 # base <dc=test,dc=com> with scope subtree # filter: (objectclass=*) # requesting: ALL # # search result search: 2 result: 0 Success # numResponses: 1 Is this the way that ACLs work in openldap? Or have I missed a configuration setting somewhere? Thanks /Cole

1 1

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

openldap-technical