openldap-bugs

openldap-bugs@openldap.org

1 participants
21001 discussions

Re: (ITS#6845) sortvals oddities - with more information
by hyc＠symas.com 28 Feb '11

28 Feb '11

aweits(a)rit.edu wrote: > Full_Name: Andrew Elble > Version: 2.4.24 / CVS Head > OS: Solaris / MacOS > URL: > Submission from: (NULL) (129.21.6.207) > Thanks for the detailed report. > > We are using sortvals on both member and memberUid. We have been > seeing duplicate member/memberUid attributes on some objects that have > been modified (as well as a lack of sorting on those attributes). It > seemed that there was a correlation between modifies to objects that > experienced deadlocks and the objects that had duplicate > member/memberUid attributes on them. We put the seqmod overlay in > place - and this reduced the number of occurrences of the issue but > did not eliminate them. > > Upon further investigation, I discovered that it was possible to > bypass the sorting behavior if the object was not created with an > instance of the attribute with sorting enabled as a part of it. > > It would seem that attr_merge() (in attr.c) should have something like this: > > if ( *a == NULL ) { > *a = attr_alloc( desc ); > if (desc->ad_type->sat_flags& SLAP_AT_SORTED_VAL) { > (*a)->a_flags |= SLAP_ATTR_SORTED_VALS; > } > } else { This is now fixed in HEAD in value.c. > Further pursuing the issue, I started to focus on the index deletion > code that was changed as a part of ITS#5183. Specifically, the > portion of code within bdb_modify_internal() (in back-bdb/modify.c) > that is commented: > > /* Move deleted values to end of array */ > > This code modifies save_attrs, which is actually apparently a pointer > to memory that resides within the cache. If a deadlock occurs, these > changes are not reverted, thereby corrupting the entry in the cache. I > replaced this code with the pre-ITS#5183 code and I am no longer able > to 'break' the object and insert duplicate member/memberUids. This is now fixed in HEAD. > I also found it surprising that the call to bdb_idl_cache_del() in > bdb_idl_delete_key() in back-bdb/idl.c occurred prior to any calls to > the database? I see what you mean. I don't think this causes any harm though. > I can answer any questions about the specifics of the environment > in which where we are seeing this - it is a somewhat difficult problem to > reproduce outside of our production environment. I'm not terribly > familiar with the code - I'm looking to see if I have collected enough > data here to open an ITS to have this fixed. (or if I'm just way off base) > > > Thanks, > > Andy > > -- -- Howard Chu CTO, Symas Corp. http://www.symas.com Director, Highland Sun http://highlandsun.com/hyc/ Chief Architect, OpenLDAP http://www.openldap.org/project/

1 0

(ITS#6847) slapd segmentation fault
by dkastens＠uos.de 28 Feb '11

28 Feb '11

Full_Name: Dirk Kastens Version: 2.4.24 OS: Red Hat Enterprise Linux Server release 5.6 (kernel 2.6.18-238.1.1.el5xen) URL: Submission from: (NULL) (2001:638:508:3d0:74e8:28f7:184a:4cdf) I updated our openldap master and two replica servers to 2.4.24 (self-compiled) last week. Saturday morning, both replicas crashed at the same time with a segementation fault: kernel: slapd[2150]: segfault at 0000000000000059 rip 00000000004fb9e7 rsp 0000000042f9e630 error 4 kernel: slapd[9451]: segfault at 0000000000000059 rip 00000000004fb9e7 rsp 000000004563f630 error 4 Running slapd with valgrind --tool=memcheck --leak-check=full gives: ==31485== LEAK SUMMARY: ==31485== definitely lost: 0 bytes in 0 blocks ==31485== indirectly lost: 0 bytes in 0 blocks ==31485== possibly lost: 110,162 bytes in 4,698 blocks ==31485== still reachable: 6,942,591 bytes in 26,974 blocks ==31485== suppressed: 0 bytes in 0 blocks ==31485== Reachable blocks (those to which a pointer was found) are not shown. ==31485== To see them, rerun with: --leak-check=full --show-reachable=yes ==31485== ==31485== For counts of detected and suppressed errors, rerun with: -v ==31485== ERROR SUMMARY: 256 errors from 256 contexts (suppressed: 16 from 7) ==31485== Warning: set address range perms: large range [0xba78000, 0x1fa7a000) (defined) ==31485== Warning: set address range perms: large range [0xba78000, 0x1fa7a000) (noaccess) ==31485== Warning: invalid file descriptor -1 in syscall read()

1 0

(ITS#6846) Deadlock in back-meta
by jorge.perez.burgos＠ericsson.com 28 Feb '11

28 Feb '11

Full_Name: Jorge Perez Burgos Version: 2.4.15 OS: SLES 10 URL: ftp://ftp.openldap.org/incoming/ Submission from: (NULL) (195.235.15.243) There is a deadlock in back-meta code, the problem is that this thread get the lock two times, the first is the one that show the trace, but previous in the method meta_back_retry it has already the lock. (gdb) bt #0 0x00002b6e1a0706a8 in __lll_mutex_lock_wait () from /lib64/libpthread.so.0 #1 0x00002b6e1a06c9fb in _L_mutex_lock_92 () from /lib64/libpthread.so.0 #2 0x00002b6e1a06c455 in pthread_mutex_lock () from /lib64/libpthread.so.0 #3 0x00000000004eb715 in meta_back_bind_op_result (op=0x2aab35d80020, rs=0x6eb6ad20, mc=0x2aab35e58be0, candidate=4, msgid=1, sendok=<value optimized out>) at bind.c:388 #4 0x00000000004eba9f in meta_back_proxy_authz_bind (mc=0x2aab35e58be0, candidate=4, op=0x2aab35d80020, rs=0x6eb6ad20, sendok=LDAP_BACK_SENDERR) at bind.c:1534 #5 0x00000000004ebc21 in meta_back_single_dobind (op=0x2aab35d80020, rs=0x6eb6ad20, mcp=0x6eb6a750, candidate=4, sendok=LDAP_BACK_SENDERR, nretries=276864893, dolock=0) at bind.c:600 #6 0x00000000004f3c4f in meta_back_retry (op=0x2aab35d80020, rs=0x6eb6ad20, mcp=0x6eb6a750, candidate=4, sendok=LDAP_BACK_SENDERR) at conn.c:721 #7 0x00000000004e9ffc in meta_back_add (op=0x2aab35d80020, rs=0x6eb6ad20) at add.c:184 #8 0x000000000044c196 in fe_op_add (op=0x2aab35d80020, rs=0x6eb6ad20) at add.c:334 #9 0x00000000004a4832 in overlay_op_walk (op=0x2aab35d80020, rs=0x6eb6ad20, which=op_add, oi=0x905850, on=0x0) at backover.c:669 #10 0x00000000004a4db5 in over_op_func (op=0x2aab35d80020, rs=0x6eb6ad20, which=op_add) at backover.c:721 #11 0x000000000044ca02 in do_add (op=0x2aab35d80020, rs=0x6eb6ad20) at add.c:194 #12 0x0000000000445bab in connection_operation (ctx=0x6eb6ae70, arg_v=<value optimized out>) at connection.c:1115 #13 0x0000000000446087 in connection_read_thread (ctx=0x6eb6ae70, argv=<value optimized out>) at connection.c:1241 #14 0x000000000052398a in ldap_int_thread_pool_wrapper (xpool=0x788e70) at tpool.c:663 #15 0x00002b6e1a06a193 in start_thread () from /lib64/libpthread.so.0 #16 0x00002b6e19814dfd in clone () from /lib64/libc.so.6

1 0

(ITS#6845) sortvals oddities - with more information
by aweits＠rit.edu 25 Feb '11

25 Feb '11

Full_Name: Andrew Elble Version: 2.4.24 / CVS Head OS: Solaris / MacOS URL: Submission from: (NULL) (129.21.6.207) We are using sortvals on both member and memberUid. We have been seeing duplicate member/memberUid attributes on some objects that have been modified (as well as a lack of sorting on those attributes). It seemed that there was a correlation between modifies to objects that experienced deadlocks and the objects that had duplicate member/memberUid attributes on them. We put the seqmod overlay in place - and this reduced the number of occurrences of the issue but did not eliminate them. Upon further investigation, I discovered that it was possible to bypass the sorting behavior if the object was not created with an instance of the attribute with sorting enabled as a part of it. It would seem that attr_merge() (in attr.c) should have something like this: if ( *a == NULL ) { *a = attr_alloc( desc ); if (desc->ad_type->sat_flags & SLAP_AT_SORTED_VAL) { (*a)->a_flags |= SLAP_ATTR_SORTED_VALS; } } else { Further pursuing the issue, I started to focus on the index deletion code that was changed as a part of ITS#5183. Specifically, the portion of code within bdb_modify_internal() (in back-bdb/modify.c) that is commented: /* Move deleted values to end of array */ This code modifies save_attrs, which is actually apparently a pointer to memory that resides within the cache. If a deadlock occurs, these changes are not reverted, thereby corrupting the entry in the cache. I replaced this code with the pre-ITS#5183 code and I am no longer able to 'break' the object and insert duplicate member/memberUids. I also found it surprising that the call to bdb_idl_cache_del() in bdb_idl_delete_key() in back-bdb/idl.c occurred prior to any calls to the database? I can answer any questions about the specifics of the environment in which where we are seeing this - it is a somewhat difficult problem to reproduce outside of our production environment. I'm not terribly familiar with the code - I'm looking to see if I have collected enough data here to open an ITS to have this fixed. (or if I'm just way off base) Thanks, Andy

1 0

(ITS#6844) AIX 6.1 | 7.1 not supported in "configure"
by joshua.taylor＠ubs.com 25 Feb '11

25 Feb '11

Full_Name: Joshua Taylor Version: 2.4.24 OS: AIX 6.1 URL: ftp://ftp.openldap.org/incoming/ Submission from: (NULL) (193.134.254.26) the configure script has no support for aix6* or aix7*, resulting in assuming no shared libraries are supported. Simply going in after each reference to "aix5*" and adding, " | aix6* | aix7* " enables shared library support. Also, support for IA64 on aix is no more. I'm not even sure there was ever a version that *actually* ran on IA64 cpus.

1 0

Re: (ITS#6841) slapd segfault error 7
by quanah＠zimbra.com 24 Feb '11

24 Feb '11

--On Tuesday, February 22, 2011 10:54 AM +0000 tecnica(a)florasul.pt wrote: > This is the output from strace -o slapd strace is useless. Please have a core file generated, with debugging symbols, and send the full backtrace. Thanks, Quanah -- Quanah Gibson-Mount Sr. Member of Technical Staff Zimbra, Inc A Division of VMware, Inc. -------------------- Zimbra :: the leader in open source messaging and collaboration

1 0

2.4.23 / Debian 6.0 i386: segfault at 811db38 ip 080c1e80 sp b2cf4b80 error 7 in slapd
by David Touzeau 24 Feb '11

24 Feb '11

Dear We have an problem using $OpenLDAP: slapd 2.4.23 (Nov 22 2010 23:39:34) $#012#011@biber:/build/buildd-openldap_2.4.23-7-i386-mi96UQ/openldap-2.4.23/debian/build/servers/slapd The same issue has been be replicated on 2.4.24 When we analyze the debug logs, the crash is performed each time we see this line : is_entry_objectclass("", "2.16.840.1.113730.3.2.6") no objectClass attribute How can we prevent this crash or fix it ? here it is a piece of logs that detect the crash. Feb 22 14:50:21 pm02 slapd[12126]: >>> dnPrettyNormal: <cn=webmail.e.florasul.pt,cn=artica,dc=my-domain,dc=com> Feb 22 14:50:21 pm02 slapd[12126]: <<< dnPrettyNormal: <cn=webmail.e.florasul.pt,cn=artica,dc=my-domain,dc=com>, <cn=webmail.e.florasul.pt,cn=artica,dc=my-domain,dc=com> Feb 22 14:50:21 pm02 slapd[12126]: bdb_dn2entry("cn=webmail.e.florasul.pt,cn=artica,dc=my-domain,dc=com") Feb 22 14:50:21 pm02 slapd[12126]: => hdb_dn2id("cn=webmail.e.florasul.pt,cn=artica,dc=my-domain,dc=com") Feb 22 14:50:21 pm02 slapd[12126]: <= hdb_dn2id: get failed: DB_NOTFOUND: No matching key/data pair found (-30988) Feb 22 14:50:21 pm02 slapd[12126]: oc_check_required entry (cn=webmail.e.florasul.pt,cn=artica,dc=my-domain,dc=com), objectClass "ArticaHostName" Feb 22 14:50:21 pm02 slapd[12126]: oc_check_allowed type "objectClass" Feb 22 14:50:21 pm02 slapd[12126]: oc_check_allowed type "cn" Feb 22 14:50:21 pm02 slapd[12126]: oc_check_allowed type "structuralObjectClass" Feb 22 14:50:21 pm02 slapd[12126]: bdb_dn2entry("cn=webmail.e.florasul.pt,cn=artica,dc=my-domain,dc=com") Feb 22 14:50:21 pm02 slapd[12126]: => hdb_dn2id("cn=webmail.e.florasul.pt,cn=artica,dc=my-domain,dc=com") Feb 22 14:50:21 pm02 slapd[12126]: <= hdb_dn2id: get failed: DB_NOTFOUND: No matching key/data pair found (-30988) Feb 22 14:50:21 pm02 slapd[12126]: is_entry_objectclass("", "2.16.840.1.113730.3.2.6") no objectClass attribute Feb 22 14:50:21 pm02 kernel: [13697.801217] slapd[13777]: segfault at 811db38 ip 080c1e80 sp b2cf4b80 error 7 in slapd[8048000+11b000] best regards

2 1

(ITS#6843) Can't slapadd cn=config LDIF with global chaining overlay
by rhafer＠suse.de 24 Feb '11

24 Feb '11

Full_Name: Ralf Haferkamp Version: HEAD OS: URL: ftp://ftp.openldap.org/incoming/ Submission from: (NULL) (89.166.251.8) slapadd-ing the following LDIF (a stripped down slapcat output) fails: ------------------------- dn: cn=config objectClass: olcGlobal cn: config dn: olcDatabase={-1}frontend,cn=config objectClass: olcDatabaseConfig objectClass: olcFrontendConfig olcDatabase: {-1}frontend dn: olcOverlay={0}chain,olcDatabase={-1}frontend,cn=config objectClass: olcOverlayConfig objectClass: olcChainConfig olcOverlay: {0}chain dn: olcDatabase={0}ldap,olcOverlay={0}chain,olcDatabase={-1}frontend,cn=config objectClass: olcChainDatabase objectClass: olcLDAPConfig olcDatabase: {0}ldap dn: olcDatabase={0}config,cn=config objectClass: olcDatabaseConfig olcDatabase: {0}config --------------------------------------- The error message is: slapadd: could not add entry dn="olcDatabase={0}config,cn=config" (line=20): Already exists The problem is the olcChainDatabase Entry. As it appears before the config-database entry config_tool_entry_put() will autocreate an olcDatabase={0}config,cn=config and after that choke up when trying to add the olcDatabase={0}config,cn=config that appears later in the LDIF file. I have a fix for this mostly ready. (Just skipping autocreation of the config-Database if child-entries of the frontendDatabase are added)

1 0

Re: (ITS#6815) Feature Request: Accesslog filter
by andrew.findlay＠skills-1st.co.uk 23 Feb '11

23 Feb '11

On Wed, Feb 23, 2011 at 02:02:17PM +0000, hyc(a)symas.com wrote: > > A URI-based restriction specification could include/exclude based on > > suffix, filter and listed attributes with a unified syntax. > > Yes... But what does the filter *mean* in > a modify op? filter on the target entry before it was modified, or after? Hmm - I was thinking of matching on the *modified attributes* (just their presence, not taking account of value). However, I can see that pre- and post-conditions might matter so: logrecord <operations> prefilter|changefilter|postfilter <LDAP URL> logignore <operations> prefilter|changefilter|postfilter <LDAP URL> ... > a search op? match the search request's filter, or filter on the search base? Both if present. Again, I might restrict the filter to matching on the presence of attributes in the filter rather than their actual value. > a compare op? Whatever works for search should be very close. Andrew -- ----------------------------------------------------------------------- | From Andrew Findlay, Skills 1st Ltd | | Consultant in large-scale systems, networks, and directory services | | http://www.skills-1st.co.uk/ +44 1628 782565 | -----------------------------------------------------------------------

1 0

Re: (ITS#6815) Feature Request: Accesslog filter
by andrew.findlay＠skills-1st.co.uk 23 Feb '11

23 Feb '11

On Wed, Feb 23, 2011 at 05:25:28AM -0800, Howard Chu wrote: > >Extending this idea slightly, would it be possible to have > >exclusions based on changes to specific attributes? The > >particular case I have in mind is where accesslog is used to > >keep a permanent audit log of changes, and ppolicy is also > >in use, resulting in one audit entry for every login > >failure. I have one site where a large proportion of the auditlog > >entries are login failures... > > Perhaps in that case, it would be simpler just to set ppolicy's mods > to be internal-only and bypass the accesslog overlay. (Currently it > does this already, if the server is a single-master replica.) There are all sorts of problems to do with ppolicy and replication... Bypassing accesslog is an interesting idea, but I can see that people who use n-failure-lockout might want the lockout events to be auditable. I wonder if there might be some benefit to giving such control on a per-operation basis. Then ppolicy might be configurable to have some automatic ops logged and others not. It could also be the start of a way to make ppolicy actions rather more global across a replication group (e.g. by putting in the ability to queue some ops for replication back to a master server). > So far you're talking about two different enhancements - the > original poster is trying to exclude a set of searches, and you're > talking about excluding modify ops. I'm not seeing any way yet to > generalize from here such that all operation types are addressed > meaningfully, and I don't want to introduce multiple special cases > to the config language. Rather than logbase <operations> <baseDN> how about logfilter <operations> <LDAP URL> where the whole clause may be repeated as many times as needed? The LDAP URL can then specify a search base, a filter, and a list of attributes of interest. Extending this to support both inclusion and exclusion clauses might be harder, as you would then need some way to give them an order of precedence. The case I mentioned before would need a rule of the form 'dont log this op if the only change concerns the atttribute pwdFailureTime'. Perhaps a simple list of rules: logrecord <operations> <LDAP URL> logignore <operations> <LDAP URL> ... and if anything is left at the end of the list then it gets logged. The filters would determine whether a log entry would be made at all, and the attribute lists would determine the content. Andrew -- ----------------------------------------------------------------------- | From Andrew Findlay, Skills 1st Ltd | | Consultant in large-scale systems, networks, and directory services | | http://www.skills-1st.co.uk/ +44 1628 782565 | -----------------------------------------------------------------------

1 0

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

openldap-bugs