[389-devel] please review: Ticket 47538 - repl-monitor.pl - HTML report does not display correct color code for lag time
https://fedorahosted.org/389//ticket/47538 https://fedorahosted.org/389/attachment/ticket/47538/0001-Ticket-47538-repl-monitor.pl-not-displaying-correct-.patch -- 389-devel mailing list 389-devel@lists.fedoraproject.org https://lists.fedoraproject.org/admin/lists/389-devel@lists.fedoraproject.org
[389-devel] Re: Please review (additional fixes): [389 Project] #48755: moving an entry could make the online init fail
On 06/17/2016 08:49 AM, Ludwig Krispenz wrote:
On 06/17/2016 05:31 PM, Noriko Hosoi wrote:
On 06/17/2016 12:17 AM, Ludwig Krispenz wrote:
Hi Noriko,
I still have a doubt on your fix. You now base the entries to be
sent only on the parentid index, but when creating a tombstone, the
entry is removed from the parentid index. So you will miss the
tombstones in the total init.
Well, I don't think a tombstone entry is removed from the parentid
index...
but in ldbm_back_delete() we remove the entryID from all inexes and if
create_tombstone_entry add it back to specific indexes
This is an example of a tombstone in id2entry.db
id 16
rdn: nsuniqueid=bd90578a-2f6311e6-9346e70f-8f7f52e7,uid=tuser0
entryid: 16
parentid: 20
This is the corresponding item on parentid.db. The entry 16 is still
there as a child of parent id 20...
# dbscan -f parentid.db -k =20 -r -n
=20 3
10 16 19
was dbcache already flushed to disk when you did run dbscan ? I get
dn:
nsuniqueid=bd76ad01-322c11e6-a5389989-c3cc84f8,cn=x,ou=People,dc=example,dc=com
entryid: 30
parentid: 2
dn: cn=y,ou=People,dc=example,dc=com
entryid: 31
parentid: 2
dbscan -f /var/lib/dirsrv/slapd-elkris3/db/example/parentid.db -k =2
-r -n
=2 1
31
I have a question about your tombstone entry... You are using entrydn
instead of entryrdn?
> *dn:*
nsuniqueid=bd76ad01-322c11e6-a5389989-c3cc84f8,cn=x,ou=People,dc=example,dc=com
> entryid: 30
> parentid: 2
Is that the default configuration for IPA? Does that mean IPA has no
move issue?
27 ldbm_back_delete( Slapi_PBlock *pb )
781 if(create_tombstone_entry)
903 if (entryrdn_get_switch()) /* subtree-rename: on */
907 /* To maintain tombstonenumsubordinates,
908 * parentid is needed for tombstone, as well. */
913 retval = index_addordel_values_sv(be,
LDBM_PARENTID_STR,
914 svals, NULL,
e->ep_id,
915 BE_INDEX_ADD, );
But in that case total update is supposed to have no change...
313 repl5_tot_run(Private_Repl_Protocol *prp)
441 if (is_entryrdn) {
// NEW BULK IMPORT
493 } else {
494 /* Original total update */
[...]
525 }
Thanks,
--noriko
Besides, the tombstone entry is found on the consumer server after
the total update/bulk import...
# ldapsearch [...] -b "dc=example,dc=com"
"(|(objectclass=nstombstone)(objectclass=ldapsubentry))" dn
dn: cn=repl keep alive 1,dc=example,dc=com
dn:
nsuniqueid=bd90578a-2f6311e6-9346e70f-8f7f52e7,uid=tuser0,ou=ou3,dc=example,dc=com
[...]
But obviously there are still issues, with which it does not pass the
IPA CI test... Continue looking into it...
Thanks,
--noriko
Ludwig
On 06/17/2016 03:05 AM, Noriko Hosoi wrote:
https://fedorahosted.org/389/ticket/48755
https://fedorahosted.org/389/attachment/ticket/48755/0001-Ticket-48755-moving-an-entry-could-make-the-online-i.3.patch
git patch file (master) -- additional fix to the server patch
https://fedorahosted.org/389/attachment/ticket/48755/0002-Ticket-48755-CI-test-test-case-for-ticket-48755.3.patch
git patch file (master) -- additional test case to the CI test patch
I also built copr build including the patch:
https://copr.fedorainfracloud.org/coprs/nhosoi/389-ds-base-1.3.5.7.f24/builds/
Martin, could it be possible to use this tentative build in the IPA
acceptance test? (I don't want to break IPA again...)
Thanks!
--noriko
--
389-devel mailing list
389-devel@lists.fedoraproject.org
https://lists.fedoraproject.org/admin/lists/389-devel@lists.fedoraproject.org
--
Red Hat GmbH,http://www.de.redhat.com/, Registered seat: Grasbrunn,
Commercial register: Amtsgericht Muenchen, HRB 153243,
Managing Directors: Charles Cachera, Michael Cunningham, Michael O'Neill, Eric
Shander
--
389-devel mailing list
389-devel@lists.fedoraproject.org
https://lists.fedoraproject.org/admin/lists/389-devel@lists.fedoraproject.org
--
389-devel mailing list
389-devel@lists.fedoraproject.org
https://lists.fedoraproject.org/admin/lists/389-devel@lists.fedoraproject.org
--
Red Hat GmbH,http://www.de.redhat.com/, Registered seat: Grasbrunn,
Commercial register: Amtsgericht Muenchen, HRB 153243,
Managing Directors: Charles Cachera, Michael Cunningham, Michael O'Neill, Eric
Shander
--
389-devel mailing list
389-devel@lists.fedoraproject.org
https://lists.fedoraproject.org/admin/lists/389-devel@lists.fedoraproject.org
--
389-devel mailing list
389-devel@lists.fedoraproject.org
https://lists.fedoraproject.org/admin/lists/389-devel@lists.fedoraproject.org
[389-devel] Re: Please review (additional fixes): [389 Project] #48755: moving an entry could make the online init fail
On 06/17/2016 12:17 AM, Ludwig Krispenz wrote: Hi Noriko, I still have a doubt on your fix. You now base the entries to be sent only on the parentid index, but when creating a tombstone, the entry is removed from the parentid index. So you will miss the tombstones in the total init. Well, I don't think a tombstone entry is removed from the parentid index... This is an example of a tombstone in id2entry.db id 16 rdn: nsuniqueid=bd90578a-2f6311e6-9346e70f-8f7f52e7,uid=tuser0 entryid: 16 parentid: 20 This is the corresponding item on parentid.db. The entry 16 is still there as a child of parent id 20... # dbscan -f parentid.db -k =20 -r -n =20 3 10 16 19 Besides, the tombstone entry is found on the consumer server after the total update/bulk import... # ldapsearch [...] -b "dc=example,dc=com" "(|(objectclass=nstombstone)(objectclass=ldapsubentry))" dn dn: cn=repl keep alive 1,dc=example,dc=com dn: nsuniqueid=bd90578a-2f6311e6-9346e70f-8f7f52e7,uid=tuser0,ou=ou3,dc=example,dc=com [...] But obviously there are still issues, with which it does not pass the IPA CI test... Continue looking into it... Thanks, --noriko Ludwig On 06/17/2016 03:05 AM, Noriko Hosoi wrote: https://fedorahosted.org/389/ticket/48755 https://fedorahosted.org/389/attachment/ticket/48755/0001-Ticket-48755-moving-an-entry-could-make-the-online-i.3.patch git patch file (master) -- additional fix to the server patch https://fedorahosted.org/389/attachment/ticket/48755/0002-Ticket-48755-CI-test-test-case-for-ticket-48755.3.patch git patch file (master) -- additional test case to the CI test patch I also built copr build including the patch: https://copr.fedorainfracloud.org/coprs/nhosoi/389-ds-base-1.3.5.7.f24/builds/ Martin, could it be possible to use this tentative build in the IPA acceptance test? (I don't want to break IPA again...) Thanks! --noriko -- 389-devel mailing list 389-devel@lists.fedoraproject.org https://lists.fedoraproject.org/admin/lists/389-devel@lists.fedoraproject.org -- Red Hat GmbH,http://www.de.redhat.com/, Registered seat: Grasbrunn, Commercial register: Amtsgericht Muenchen, HRB 153243, Managing Directors: Charles Cachera, Michael Cunningham, Michael O'Neill, Eric Shander -- 389-devel mailing list 389-devel@lists.fedoraproject.org https://lists.fedoraproject.org/admin/lists/389-devel@lists.fedoraproject.org -- 389-devel mailing list 389-devel@lists.fedoraproject.org https://lists.fedoraproject.org/admin/lists/389-devel@lists.fedoraproject.org
[389-devel] Re: Please review: ticket #48366 - proxyauth support does not work when bound as directory manager
I added some clarifications to the ticket and here is a lib389 test case https://fedorahosted.org/389/attachment/ticket/48366/0001-add-testcase-for-ticket-48366-proxyauth-for-root.patch On 02/16/2016 03:35 PM, Ludwig Krispenz wrote: https://fedorahosted.org/389/ticket/48366 https://fedorahosted.org/389/attachment/ticket/48366/0001-Ticket-48366-proxyauth-does-not-work-bound-as-direct.patch -- 389-devel mailing list 389-devel@%(host_name)s http://lists.fedoraproject.org/admin/lists/389-devel@lists.fedoraproject.org -- Red Hat GmbH, http://www.de.redhat.com/, Registered seat: Grasbrunn, Commercial register: Amtsgericht Muenchen, HRB 153243, Managing Directors: Charles Cachera, Michael Cunningham, Michael O'Neill, Eric Shander -- 389-devel mailing list 389-devel@lists.fedoraproject.org https://lists.fedoraproject.org/admin/lists/389-devel@lists.fedoraproject.org
[389-devel] Re: Please review (additional fixes): [389 Project] #48755: moving an entry could make the online init fail
On 06/17/2016 02:36 PM, Martin Babinsky wrote: On 06/17/2016 08:42 AM, Martin Babinsky wrote: On 06/17/2016 03:05 AM, Noriko Hosoi wrote: https://fedorahosted.org/389/ticket/48755 https://fedorahosted.org/389/attachment/ticket/48755/0001-Ticket-48755-moving-an-entry-could-make-the-online-i.3.patch git patch file (master) -- additional fix to the server patch https://fedorahosted.org/389/attachment/ticket/48755/0002-Ticket-48755-CI-test-test-case-for-ticket-48755.3.patch git patch file (master) -- additional test case to the CI test patch I also built copr build including the patch: https://copr.fedorainfracloud.org/coprs/nhosoi/389-ds-base-1.3.5.7.f24/builds/ Martin, could it be possible to use this tentative build in the IPA acceptance test? (I don't want to break IPA again...) Thanks! --noriko Hi Noriko, I will run some of our replication CI tests with your copr build and get back with results. Our simple-replication CI test failed with your copr build and master branch freeipa (built today). The previous DS build (389-ds-base-1.3.5.6-1.fc24.x86_64) worked just fine. These are the errors/access logs from the failed test: https://paste.fedoraproject.org/380434/15539814/ https://paste.fedoraproject.org/380435/61554061/ these seem to be from the replica, do you have the logs from master ? I will do more investigation next week as we have to prioritize tasks heavily due to impending deadlines. -- Red Hat GmbH, http://www.de.redhat.com/, Registered seat: Grasbrunn, Commercial register: Amtsgericht Muenchen, HRB 153243, Managing Directors: Charles Cachera, Michael Cunningham, Michael O'Neill, Eric Shander -- 389-devel mailing list 389-devel@lists.fedoraproject.org https://lists.fedoraproject.org/admin/lists/389-devel@lists.fedoraproject.org
[389-devel] Re: Please review (additional fixes): [389 Project] #48755: moving an entry could make the online init fail
On 06/17/2016 03:05 AM, Noriko Hosoi wrote: https://fedorahosted.org/389/ticket/48755 https://fedorahosted.org/389/attachment/ticket/48755/0001-Ticket-48755-moving-an-entry-could-make-the-online-i.3.patch git patch file (master) -- additional fix to the server patch https://fedorahosted.org/389/attachment/ticket/48755/0002-Ticket-48755-CI-test-test-case-for-ticket-48755.3.patch git patch file (master) -- additional test case to the CI test patch I also built copr build including the patch: https://copr.fedorainfracloud.org/coprs/nhosoi/389-ds-base-1.3.5.7.f24/builds/ Martin, could it be possible to use this tentative build in the IPA acceptance test? (I don't want to break IPA again...) Thanks! --noriko Hi Noriko, I will run some of our replication CI tests with your copr build and get back with results. -- Martin^3 Babinsky -- 389-devel mailing list 389-devel@lists.fedoraproject.org https://lists.fedoraproject.org/admin/lists/389-devel@lists.fedoraproject.org
