openldap-bugs September 2007

openldap-bugs@openldap.org

33 participants
179 discussions

Re: (ITS#5134) dbconfig vs. DB_CONFIG
by ando＠sys-net.it 13 Sep '07

13 Sep '07

hyc(a)symas.com wrote: >> Reading the man page, I thought that the directives don't modify >> DB_CONFIG, but were applied at slapd's startup. > > For slapd.conf, the directives can only take effect after they are written to > DB_CONFIG. They will only be written to DB_CONFIG at startup time, and only if > no such file already existed. > > The main reason this directive was added was for the benefit of cn=config. > Using it in slapd.conf is rather pointless. Once you start managing slapd > through cn=config, you are expected to stop editing DB_CONFIG manually. I didn't mention that the whole discussion started because Debian seems to ship with both a default DB_CONFIG installed in the database directory, and the very same directives echoed in slapd.conf. People find the dbconfig directives, modify them and complain because they don't take effect. And, I couldn't disambiguate the man page because I didn't find it clear enough about this point. p. Ing. Pierangelo Masarati OpenLDAP Core Team SysNet s.r.l. via Dossi, 8 - 27100 Pavia - ITALIA http://www.sys-net.it --------------------------------------- Office: +39 02 23998309 Mobile: +39 333 4963172 Email: pierangelo.masarati(a)sys-net.it ---------------------------------------

1 0

Re: (ITS#5134) dbconfig vs. DB_CONFIG
by rra＠stanford.edu 13 Sep '07

13 Sep '07

hyc(a)symas.com writes: > Raphaël Ouazana-Sustowski wrote: >> Thank you very much for the clarification. >> Just for the records, it seems that Debian packages (and maybe others) >> use dbconfig directives without warning users that they should not >> modify these parameters here... > Thanks for that bit of information, I was not aware of that. For the > record, Debian packagers have never asked us whether what they're doing > makes any bit of sense. Well, that's not exactly true, given that I have in fact talked with Quanah about this and he recommended setting everything in slapd.conf instead of using a separate DB_CONFIG file, although the code in the Debian package predates that. Certainly there should be a comment explaining that changes there only take effect if the DB_CONFIG file doesn't exist. I'll file a bug against the Debian package for whenever someone gets a chance to upload a new version. (I wish I had more time to work on this.) -- Russ Allbery (rra(a)stanford.edu) <http://www.eyrie.org/~eagle/>

1 0

Re: (ITS#5134) dbconfig vs. DB_CONFIG
by ghenry＠suretecsystems.com 13 Sep '07

13 Sep '07

<quote who="Howard Chu"> > Gavin Henry wrote: >> <quote who="Howard Chu"> >>> ghenry(a)suretecsystems.com wrote: >>>> I would say since dbconfig directives only work on a restart of slapd, >>>> and >>>> that if a DB_CONFIG is present, the DB_CONFIG should *always* take >>>> precedence. If dbconfig changes could make a difference via >>>> back-config, >>>> which they can't without a restart, then dbconfig stuff should >>>> override >>>> any DB_CONFIG. >>> dbconfig behavior via cn=config is fully dynamic, and thus quite >>> different >>> from >>> slapd.conf. Changes to dbconfig via cn=config take effect immediately >>> and >>> will >>> completely rewrite the DB_CONFIG file. >> >> Understood, but a restart would still be needed to make the db env >> honour >> DB_CONFIG, right? > > NO. "Fully dynamic" and "take effect immediately" means exactly that. I > did not > say "the file is written immediately but there's no effect until the > server is > restarted." Immediate effect means immediate effect. Do we really have to > say > everything twice or thrice here before people get it? Ok, I get it. I didn't know the db env was reopened when needed with cn=config. Now I do. I'll make sure this is in the docs. I'll try to make it crystal clear! P.S. What does thrice mean? ;-)

1 0

Re: (ITS#5133) Synchronous replication on slave doesn't notice lost network connection
by ando＠sys-net.it 13 Sep '07

13 Sep '07

audrius.valunas(a)teo.lt wrote: > There is synchronous replication between mastyer and slave. When network > connectivity problems occur master closes tcp connection but slave doesn't > notice those problems, it still has tcp connection open, but in real it is not > receiving updates any more. > I think that can be solved adding some ack from slave because sending on such a > socket would fail and force slave to retry connection. Well, this should already be taken into consideration by SO_KEEPALIVE, which is always set when available on all connections. I concur that it usually requires quite a long time before a connection is actually checked (usually more than 2 hours), so some better policy could be put in place. p. Ing. Pierangelo Masarati OpenLDAP Core Team SysNet s.r.l. via Dossi, 8 - 27100 Pavia - ITALIA http://www.sys-net.it --------------------------------------- Office: +39 02 23998309 Mobile: +39 333 4963172 Email: pierangelo.masarati(a)sys-net.it ---------------------------------------

1 0

Re: (ITS#5134) dbconfig vs. DB_CONFIG
by hyc＠symas.com 13 Sep '07

13 Sep '07

Raphaël Ouazana-Sustowski wrote: > Le Jeu 13 septembre 2007 22:40, Howard Chu a écrit : >> Raphaël Ouazana-Sustowski wrote: >>> Hi, >>> >>> Le Jeu 13 septembre 2007 21:29, hyc(a)symas.com a écrit : >>>> There is nothing ambiguous here. When slapd.conf is being used, the >>>> dbconfig >>>> directive *only* takes effect *if* no DB_CONFIG file exists. If one >>>> exists, the >>>> dbconfig directives are ignored. >>> So if someone first create its DB_CONFIG by putting dbconfig directives >>> in >>> slapd.conf, and then modify these directives, the modification has no >>> effect ? >> Correct. >> >>> Reading the man page, I thought that the directives don't modify >>> DB_CONFIG, but were applied at slapd's startup. >> For slapd.conf, the directives can only take effect after they are written >> to >> DB_CONFIG. They will only be written to DB_CONFIG at startup time, and >> only if >> no such file already existed. >> >> The main reason this directive was added was for the benefit of cn=config. >> Using it in slapd.conf is rather pointless. Once you start managing slapd >> through cn=config, you are expected to stop editing DB_CONFIG manually. > > Thank you very much for the clarification. > > Just for the records, it seems that Debian packages (and maybe others) use > dbconfig directives without warning users that they should not modify > these parameters here... Thanks for that bit of information, I was not aware of that. For the record, Debian packagers have never asked us whether what they're doing makes any bit of sense. I believe it is the packager's responsibility to proactively check these things out; certainly we (OpenLDAP Project) cannot go out and poll every packager. -- -- Howard Chu Chief Architect, 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#5134) dbconfig vs. DB_CONFIG
by hyc＠symas.com 13 Sep '07

13 Sep '07

Pierangelo Masarati wrote: > Howard Chu wrote: >> "If (x) then (y)" is the complete definition of the behavior. You are >> looking for explanations for "If (!x)", but it should be well understood >> that such cases are irrelevant given this definition. > > That's actually an if (x) and (!y) then (z). What is missing is if (x) > and (y) (which happens to be (!z), but it could be anything...) Yes, but "anything" would require extra work. (Kinda like #pragma in gcc 1.x would start up NetHack, since the C spec said the behavior was totally implementation defined.) Nor does it make sense for any other outcome to exist. E.g. with your example, you're supposing the possibility that "if (x) and (y) then (z)" but then the outcome is identical for both "if (x) and (!y)" and "if (x) and (y)" which makes y completely irrelevant. Again, nobody would go to the bother of jumping thru the hoops for these extra conditions if they are ultimately irrelevant. -- -- Howard Chu Chief Architect, 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#5134) dbconfig vs. DB_CONFIG
by raphael.ouazana＠linagora.com 13 Sep '07

13 Sep '07

Le Jeu 13 septembre 2007 22:40, Howard Chu a écrit : > Raphaël Ouazana-Sustowski wrote: >> Hi, >> >> Le Jeu 13 septembre 2007 21:29, hyc(a)symas.com a écrit : >>> There is nothing ambiguous here. When slapd.conf is being used, the >>> dbconfig >>> directive *only* takes effect *if* no DB_CONFIG file exists. If one >>> exists, the >>> dbconfig directives are ignored. >> >> So if someone first create its DB_CONFIG by putting dbconfig directives >> in >> slapd.conf, and then modify these directives, the modification has no >> effect ? > > Correct. > >> Reading the man page, I thought that the directives don't modify >> DB_CONFIG, but were applied at slapd's startup. > > For slapd.conf, the directives can only take effect after they are written > to > DB_CONFIG. They will only be written to DB_CONFIG at startup time, and > only if > no such file already existed. > > The main reason this directive was added was for the benefit of cn=config. > Using it in slapd.conf is rather pointless. Once you start managing slapd > through cn=config, you are expected to stop editing DB_CONFIG manually. Thank you very much for the clarification. Just for the records, it seems that Debian packages (and maybe others) use dbconfig directives without warning users that they should not modify these parameters here... Regards, Raphael Ouazana.

1 0

Re: (ITS#4860) Sets' enhancement
by ando＠sys-net.it 13 Sep '07

13 Sep '07

Should be fixed now in HEAD/re24/re23. Please test. p. Ing. Pierangelo Masarati OpenLDAP Core Team SysNet s.r.l. via Dossi, 8 - 27100 Pavia - ITALIA http://www.sys-net.it --------------------------------------- Office: +39 02 23998309 Mobile: +39 333 4963172 Email: pierangelo.masarati(a)sys-net.it ---------------------------------------

1 0

Re: (ITS#5134) dbconfig vs. DB_CONFIG
by hyc＠symas.com 13 Sep '07

13 Sep '07

Raphaël Ouazana-Sustowski wrote: > Hi, > > Le Jeu 13 septembre 2007 21:29, hyc(a)symas.com a écrit : >> There is nothing ambiguous here. When slapd.conf is being used, the >> dbconfig >> directive *only* takes effect *if* no DB_CONFIG file exists. If one >> exists, the >> dbconfig directives are ignored. > > So if someone first create its DB_CONFIG by putting dbconfig directives in > slapd.conf, and then modify these directives, the modification has no > effect ? Correct. > Reading the man page, I thought that the directives don't modify > DB_CONFIG, but were applied at slapd's startup. For slapd.conf, the directives can only take effect after they are written to DB_CONFIG. They will only be written to DB_CONFIG at startup time, and only if no such file already existed. The main reason this directive was added was for the benefit of cn=config. Using it in slapd.conf is rather pointless. Once you start managing slapd through cn=config, you are expected to stop editing DB_CONFIG manually. -- -- Howard Chu Chief Architect, 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#5134) dbconfig vs. DB_CONFIG
by ando＠sys-net.it 13 Sep '07

13 Sep '07

Howard Chu wrote: > Pierangelo Masarati wrote: >> hyc(a)symas.com wrote: >> >>> There is nothing ambiguous here. When slapd.conf is being used, the >>> dbconfig directive *only* takes effect *if* no DB_CONFIG file exists. >>> If one exists, the dbconfig directives are ignored. >> >> I insist: I don't see it so clear, and I was pointed to it by other >> people that didn't find it clear as well. I'm not questioning what the >> code does, just this very bit of the man page. > > "If (x) then (y)" is the complete definition of the behavior. You are > looking for explanations for "If (!x)", but it should be well understood > that such cases are irrelevant given this definition. That's actually an if (x) and (!y) then (z). What is missing is if (x) and (y) (which happens to be (!z), but it could be anything...) p. Ing. Pierangelo Masarati OpenLDAP Core Team SysNet s.r.l. via Dossi, 8 - 27100 Pavia - ITALIA http://www.sys-net.it --------------------------------------- Office: +39 02 23998309 Mobile: +39 333 4963172 Email: pierangelo.masarati(a)sys-net.it ---------------------------------------

1 0

← Newer
1
...
6
7
8
9
10
11
12
...
18
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 September 2007