Full_Name: Balint Reczey Version: master OS: Linux URL: http://rbalint.cs.bme.hu/0001-slapcat-retry-a-few-times-when-retrieving-an-entry-f.patch Submission from: (NULL) (84.0.168.17) From: Axel Beckert <beckert@phys.ethz.ch> To: Debian Bug Tracking System <submit@bugs.debian.org> Subject: slapd: slapcat output truncated every now and then Date: Tue, 15 May 2012 18:33:57 +0200 Package: slapd Severity: grave Version: 2.4.23-7.2 Justification: Causes data loss Every now and then slapcat's output does not contain the whole content of the LDAP but is truncated at some LDIF entry border (i.e. all printed LDIF records seem complete). To reproduce run "while sleep 1; do slapcat | wc; done" on a moderately busy LDAP server (writes every few seconds to minutes; about 10000 entries) while slapd is running and notice the occasionally occurring huge change in wc's printed values, e.g.: # while sleep 1; do slapcat | wc; done 471698 1015498 15336677 471698 1015498 15336677 471698 1015498 15336677 ... 471698 1015497 15336635 471698 1015497 15336635 281829 606820 8632165 <-- 471698 1015497 15336635 471698 1015497 15336635 471698 1015497 15336635 308627 664573 9476751 <-- 471698 1015497 15336635 471698 1015497 15336635 471698 1015497 15336637 Please see email thread starting at http://www.openldap.org/lists/openldap-technical/201301/msg00195.html The attached modifications to OpenLDAP Software are subject to the following notice: Copyright 2013 Balint Reczey <balint@balintreczey.hu> Redistribution and use in source and binary forms, with or without modification, are permitted only as authorized by the OpenLDAP Public License.
--On Friday, January 25, 2013 10:26 PM +0000 balint@balintreczey.hu wrote: > Full_Name: Balint Reczey > Version: master > OS: Linux > URL: ftp://ftp.openldap.org/incoming/ > Submission from: (NULL) (84.0.168.17) Hi Balint, The patch as submitted is rejected, due to the fact that it changes slapadd, rather than fixing back-bdb/hdb, which is where the actual problem occurs. I.e., your change is too broad. Feel free to submit an updated patch correctly updating the back-bdb/hdb code rather than slapadd. 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
--On Sunday, January 27, 2013 6:09 PM +0000 quanah@zimbra.com wrote: > --On Friday, January 25, 2013 10:26 PM +0000 balint@balintreczey.hu wrote: > >> Full_Name: Balint Reczey >> Version: master >> OS: Linux >> URL: ftp://ftp.openldap.org/incoming/ >> Submission from: (NULL) (84.0.168.17) > > Hi Balint, > > The patch as submitted is rejected, due to the fact that it changes > slapadd, rather than fixing back-bdb/hdb, which is where the actual > problem occurs. I.e., your change is too broad. Feel free to submit > an updated patch correctly updating the back-bdb/hdb code rather than > slapadd. s/slapadd/slapcat/g ;) --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
2013/1/27 Quanah Gibson-Mount <quanah@zimbra.com>: > --On Friday, January 25, 2013 10:26 PM +0000 balint@balintreczey.hu wrote: > >> Full_Name: Balint Reczey >> Version: master >> OS: Linux >> URL: ftp://ftp.openldap.org/incoming/ >> Submission from: (NULL) (84.0.168.17) > > > Hi Balint, > > The patch as submitted is rejected, due to the fact that it changes slapadd, > rather than fixing back-bdb/hdb, which is where the actual problem occurs. > I.e., your change is too broad. Feel free to submit an updated patch > correctly updating the back-bdb/hdb code rather than slapadd. > > Thanks, > Quanah > Hi Quanah, I'll let OpenLDAP developers fix the problem where they find it appropriate instead. My intention was adding a retry logic where there was a skip and continue logic already, in slapcat. Cheers, Balint
--On Sunday, January 27, 2013 10:03 PM +0100 Bálint Réczey <balint@balintreczey.hu> wrote: > I'll let OpenLDAP developers fix the problem where they find it > appropriate instead. > My intention was adding a retry logic where there was a skip and > continue logic already, in slapcat. Hi Balint, I understand your intent, I'm simply noting it was the wrong approach. Supplying a patch that fixes the actual issue, if you choose to do so, simply means it will have a greater chance of being fixed sooner than later. I.e., it is not a high priority issue at this time for the openldap devs. --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
I've committed an alternate fix for this to git master, filed against ITS#6365. Please test and let us know your results, thanks. The patch is in commit 853b9d1335d27e280751e9cfb8ca6b5356ffec73 -- -- Howard Chu CTO, Symas Corp. http://www.symas.com Director, Highland Sun http://highlandsun.com/hyc/ Chief Architect, OpenLDAP http://www.openldap.org/project/
Hi Howard, Thank you! I have updated the Debian bug report [1]. Cheers, Balint [1]: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=673038 2013/2/7 Howard Chu <hyc@symas.com>: > I've committed an alternate fix for this to git master, filed against > ITS#6365. Please test and let us know your results, thanks. > The patch is in commit 853b9d1335d27e280751e9cfb8ca6b5356ffec73 > > -- > -- Howard Chu > CTO, Symas Corp. http://www.symas.com > Director, Highland Sun http://highlandsun.com/hyc/ > Chief Architect, OpenLDAP http://www.openldap.org/project/
changed notes changed state Open to Closed
fixed as ITS#6365