openldap-bugs April 2008

openldap-bugs@openldap.org

45 participants
192 discussions

Start a nNew thread

Re: (ITS#5451) syncprov_checkpoint deadlock bugfix?
by h.b.furuseth＠usit.uio.no 07 Apr '08

07 Apr '08

Can you post your master and slave slapd.conf files, after removing passwords? -- Hallvard

1 0

Re: (ITS#5456) inaccurate interpretation of LDIF in back perl
by Uwe.Werler＠o3si.de 07 Apr '08

07 Apr '08

Hello Howard, in my last message (see below) the question was not wether is an entry starting with a blank line or not or what is a correct ldif. My question was if back_perl is only accepting "ldif-like" entries starting with dn-spec. If I return an entry like this: version: 1 dn: ou=foo,dc=bar ou: foo objectClass: organizationalUnit objectClass: top description: users from foo or with SEP after version-spec version: 1 dn: ou=foo,dc=bar ou: foo objectClass: organizationalUnit objectClass: top description: users from foo then the debug output from slapd is: "str2entry: entry -1 has no dn" That means that back_perl doesn't accept ldif files but only "ldif-like" entries without version-spec. If this is intended than I suggest to explain this in the man page or at least in the example shipped with back_perl. Thank You for Your attention. Regards Uwe Am Sonntag, den 06.04.2008, 12:41 -0700 schrieb Howard Chu: > Please do not email me directly unless you are a paying client of my company. > This thread started on the ITS. It should remain on the ITS. > > I already explained to you what the RFC spec says about valid LDIF. Valid LDIF > does not start with an empty line. There is nothing further to explain. > > Uwe Werler wrote: > > Dear Howard, > > > > after some testing again I determined that back_perl doesn't accept > > ldif_entries which are starting with an empty line. Then slapd tells me > > "str2entry: entry -1 has no dn". > > > > Is this behavior correct? If so it will be good to explain in the man > > page or the example that an entry should start with a dn line and > > shouldn't contain any empty lines nor the version information at the > > first line. > > > > I think that my mistake was to use sample code I've found like this: > > > > sub search > > { > > > > my $this = shift; > > my($base, $scope, $deref, $sizeLim, $timeLim, $filterStr, $attrOnly, @attrs ) = @_; > > if (! defined($this->{'LDAP'})) { > > $ldap = Net::LDAP->new( 'localhost' ) ; > > $this->{'LDAP'} = $ldap; > > } > > $mesg=$this->{'LDAP'}->search(base=>$base, > > scope=>$scope, > > deref=>$deref, > > sizelimit =>$sizeLim, > > timelimit =>$timeLim, > > filter =>$filterStr, > > attrs=> @attrs ); > > > > my @list = (); > > > > > > foreach $entry ($mesg->all_entries) { > > my $data; > > $LDIF = new IO::Scalar \$data; > > $dump=Net::LDAP::LDIF->new($LDIF,"w"); > > $dump->write($entry); > > $LDIF->close; > > push @list,$data; > > } > > return ( 0 , @list); > > > > } > > > > Because the entries returned by Net::LDAP::LDIF are starting with a leading SEP it won't work as is in back_perl. > > > > Thank You very much for Your attention. > > > > Regards Uwe > > > > > >

1 0

Re: (ITS#5451) syncprov_checkpoint deadlock bugfix?
by rein＠basefarm.no 07 Apr '08

07 Apr '08

rein(a)basefarm.no skrev: >> We have seen deadlock/hang situations in our master server today, which looks >> like a deadlock caused by the si_csn_rwlock lock being held while >> syncprov_checkpoint() is running. The patch at the end should (I hope) fix >> this. > > Hm, I looks as if I was a bit to quick here, it just hung got again :-(. > And this time without anyone competing for this lock, only the lock on > the glue suffix entry. I wonder if it can be the upgrade from db 4.6.18 > to 4.6.21.1 I dit yesterday that is the real problem. I'll try to > downgrade and see if that helps. And it didn't :-( The problem seem to be that something readlocks our glue suffix entry before forgetting about the lock. Which quickly causes the entire server to deadlock when the writelock required to update the contextCSN in the suffix entry locks out all the readers. So far the problem seem to be triggered by someone attempting to modify an entry in a subordinate syncrepl consumer backend that results in a referral to the backend master. But I haven't had very much time to look into this problem yet, so I'm still on very thin ice here. I'll return with a new ITS when I have found out more. We are currently running with a workaround that simply grants the writelocks on the glue suffix entry without actually doing it. As the glue entry is the only entry in that database it should be pretty safe, and a potential corruption of the database is not any big problem. > Please put this case on hold. Sorry! It currently looks as this patch addresses the symptom and not the real problem, although I'm not sure what could happen if a checkpoint is triggered while the suffix entry is locked by another thread. I do believe it should be considered if not a bug so an enhancement, as holding locks for as short time as possible is always a good thing. You'll have to choose whether to close this ITS or use the patch. Rein

1 0

Re: (ITS#5454) syncrepl refreshAndPersist stops receiving
by hyc＠symas.com 07 Apr '08

07 Apr '08

Rein Tollevik wrote: > Howard Chu skrev: >> How about just moving the enable to after the runqueue manipulation is >> done? > > That would still leave a glitch in the window open, unless the initial > trylock is changed to a regular lock as your checking message suggests. > > But, doing that could result in all the threads ending up waiting for > the lock if do_syncrepl of a refresh only replication is configured to > run too often. I assume the trylock is there to avoid this? Yes, but I don't think that's a problem any more. If the scheduled task is still running when the next interval hits, the daemon task will just reschedule; it won't fire off a new thread for the task. >> It also occurs to me that we probably don't even need to manipulate the >> slapd runqueue in persist mode, when si->si_conn is already set. I.e., >> in that case we can only have gotten here because of a listener event, >> and not because of a runqueue schedule. > > That is probably true. Right, my patch avoids that now. Having looked at it again I think replacing the trylock with a regular lock is fine. -- -- 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

Re: (ITS#5454) syncrepl refreshAndPersist stops receiving
by rein＠basefarm.no 07 Apr '08

07 Apr '08

Howard Chu skrev: > rein(a)basefarm.no wrote: >> Hm, I hope I have found the race condition that causes this :-) I'm now >> running with the patch at the end to see if that solves it, only time >> will tell.. And we haven't seen any replication problems since this was implemented friday afternoon :-). But it is still a bit too early to conclude that it has been fixed. >> The race is that between the time selecting on the syncrepl socket is >> enabled by the call to connection_client_enable() and the release of the >> si_mutex a new message may arrive. If so, the next call to do_syncrepl >> may fail in its attempt to trylock the mutex and no-one will re-enable >> selecting on it again. My patch delays enabling of the socket until the >> mutex has been released, which looks safe to me. Or can the access to >> si->si_conn without a lock be a problem? > > How about just moving the enable to after the runqueue manipulation is > done? That would still leave a glitch in the window open, unless the initial trylock is changed to a regular lock as your checking message suggests. But, doing that could result in all the threads ending up waiting for the lock if do_syncrepl of a refresh only replication is configured to run too often. I assume the trylock is there to avoid this? What about starting out with deferring the next scheduled call, as a scheduled call is not wanted until the first has completed anyhow? And a rescedule is already being done before do_syncrepl finishes. Deferring should not be necessary if rtask->next_sched.tv_sec is zero, i.e it has already been deferred. That would remove these calls for all the cases when do_syncrepl is called as a result of a new message on a persistent syncrepl connection. > Just need to be sure that do_syncrepl() isn't entered again before > si->si_conn gets initialized. I'm not quite sure what you mean here. As I see it, the problem is that do_syncrepl() must not be called after si->si_conn is enabled and the lock is still held. > It also occurs to me that we probably don't even need to manipulate the > slapd runqueue in persist mode, when si->si_conn is already set. I.e., > in that case we can only have gotten here because of a listener event, > and not because of a runqueue schedule. That is probably true. Rein

1 0

Re: (ITS#5456) inaccurate interpretation of LDIF in back perl
by h.b.furuseth＠usit.uio.no 07 Apr '08

07 Apr '08

I wrote: > You two seem to be saying the same thing: An entry should end with a > line terminator, but not with several line terminators (which give > blank lines). Duh, sorry, I didn't see you said it _begins_ with a newline. Anyway, Net::LDAP::LDIF is doing the right thing: When you write multiple entries to the same Net::LDAP::LDIF object, it puts an empty line between them. -- Hallvard

1 0

Re: (ITS#5404) back-ldap fails intermittently with 80 Internal (implementation specific) error
by ghenry＠OpenLDAP.org 06 Apr '08

06 Apr '08

toby(a)inf.ed.ac.uk wrote: >> Please test and report; >> >> servers/slapd/back-ldap/bind.c >> new revision: 1.244; previous revision: 1.243 >> >> The patch applies straightforwardly to re23 only if the first chunk is >> omitted. > > Been running for 3+ weeks now on various test machines with no > recurrence of error, so I'd consider this resolved. > > Many thanks > Toby > > > Thanks for the followup. Gavin. -- Kind Regards, Gavin Henry. OpenLDAP Engineering Team. E ghenry(a)OpenLDAP.org Community developed LDAP software. http://www.openldap.org/project/

1 0

Re: (ITS#5456) inaccurate interpretation of LDIF in back perl
by hyc＠symas.com 06 Apr '08

06 Apr '08

Please do not email me directly unless you are a paying client of my company. This thread started on the ITS. It should remain on the ITS. I already explained to you what the RFC spec says about valid LDIF. Valid LDIF does not start with an empty line. There is nothing further to explain. Uwe Werler wrote: > Dear Howard, > > after some testing again I determined that back_perl doesn't accept > ldif_entries which are starting with an empty line. Then slapd tells me > "str2entry: entry -1 has no dn". > > Is this behavior correct? If so it will be good to explain in the man > page or the example that an entry should start with a dn line and > shouldn't contain any empty lines nor the version information at the > first line. > > I think that my mistake was to use sample code I've found like this: > > sub search > { > > my $this = shift; > my($base, $scope, $deref, $sizeLim, $timeLim, $filterStr, $attrOnly, @attrs ) = @_; > if (! defined($this->{'LDAP'})) { > $ldap = Net::LDAP->new( 'localhost' ) ; > $this->{'LDAP'} = $ldap; > } > $mesg=$this->{'LDAP'}->search(base=>$base, > scope=>$scope, > deref=>$deref, > sizelimit =>$sizeLim, > timelimit =>$timeLim, > filter =>$filterStr, > attrs=> @attrs ); > > my @list = (); > > > foreach $entry ($mesg->all_entries) { > my $data; > $LDIF = new IO::Scalar \$data; > $dump=Net::LDAP::LDIF->new($LDIF,"w"); > $dump->write($entry); > $LDIF->close; > push @list,$data; > } > return ( 0 , @list); > > } > > Because the entries returned by Net::LDAP::LDIF are starting with a leading SEP it won't work as is in back_perl. > > Thank You very much for Your attention. > > Regards Uwe > > -- -- 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

Re: (ITS#5456) inaccurate interpretation of LDIF in back perl
by h.b.furuseth＠usit.uio.no 05 Apr '08

05 Apr '08

hyc(a)symas.com writes: > uw(a)o3si.de wrote: >> The database backend "back_perl" expects LDIF entries beginning with >> "dn:" and ending with an empty line. (...) > > Wrong. (...) all records must end with LF or CR/LF. You two seem to be saying the same thing: An entry should end with a line terminator, but not with several line terminators (which give blank lines). >> After some testing with back_perl I noticed that the LDIF entries >> produced by the perl module Net::LDAP::LDIF were not acceptet because >> they are beginning with an empty line and ending only with a simple >> CR charakter. If you really mean CR (\r), that's a Net::LDAP::LDIF bug. However Net::LDAP::LDIF on my host correctly ends an entry with LF (\n). And back-perl (at least the search operation) accepts entries ending with both zero, one and two '\n's, as well as with '\r'. There was however an entry parsing bug (ITS#4852) in 2.3.32 which was fixed in 2.3.35. I don't see how it'd be related, but I suggest you upgrade anyway. Then if you still have trouble, try a simple perl script which does something like sub search { 0, "dn: o=foo\no: foo\nobjectClass: organization\n" } instead of going via a module like Net::LDAP::LDIF, so you can be quite sure the problem is what you think it is. -- Hallvard

1 0

Re: (ITS#5456) inaccurate interpretation of LDIF in back perl
by hyc＠symas.com 05 Apr '08

05 Apr '08

uw(a)o3si.de wrote: > Full_Name: Uwe Werler > Version: 2.3.32 > OS: SLES10 SP1 Kernel 2.6.16.54-0.2.3-xen > URL: ftp://ftp.openldap.org/incoming/ > Submission from: (NULL) (155.56.68.221) > > > The database backend "back_perl" expects LDIF entries beginning with "dn:" and > ending with an empty line. RFC2849 only describes that an entry is everything > except an empty line and a line not beginning (after a CR/LF) with a single > white space character. Wrong. The grammar in RFC2849 quite explicitly states: ldif-file = ldif-content / ldif-changes ldif-content = version-spec 1*(1*SEP ldif-attrval-record) ldif-changes = version-spec 1*(1*SEP ldif-change-record) ldif-attrval-record = dn-spec SEP 1*attrval-spec ldif-change-record = dn-spec SEP *control changerecord That means an LDIF file must start with 0 or 1 version-specs, then a dn-spec. And, all records must end with LF or CR/LF. > After some testing with back_perl I noticed that the LDIF entries produced by > the perl module Net::LDAP::LDIF were not acceptet because they are beginning > with an empty line and ending only with a simple CR charakter. > > This behavior breaks RFC2849. After changing LDIF.pm from that module to produce > the CR after the entry back_perl accepts this entries too. > > IMHO this is a misbehavior so I'm very glad if You'll take a look at this > issue. Apparently it Net::LDAP::LDIF is broken. Since that is not a piece of OpenLDAP Software I don't see what we can do about it. This ITS will be closed. Contact the author of Net::LDAP::LDIF to pursue this further. -- -- 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

← Newer
1
...
13
14
15
16
17
18
19
20
Older →

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

openldap-bugs April 2008