Skip to Content.
Sympa Menu

grouper-users - Re: [grouper-users] PSP problem after folder rename

Subject: Grouper Users - Open Discussion List

List archive

Re: [grouper-users] PSP problem after folder rename


Chronological Thread 
  • From: "Bryan E. Wooten" <>
  • To: Tim Darby <>, Tom Zeller <>
  • Cc: "" <>
  • Subject: Re: [grouper-users] PSP problem after folder rename
  • Date: Sat, 18 May 2013 00:18:25 +0000
  • Accept-language: en-US
  • Authentication-results: sfpop-ironport03.merit.edu; dkim=neutral (message not signed) header.i=none

Tim,

Glad to hear that I am not the only one that makes minor mistakes causing way too much grief for myself and others.

When I ran into issues there were over 100k change log entries, it took a long time to run a test.

-Bryan

From: Tim Darby <>
Date: Friday, May 17, 2013 3:49 PM
To: Tom Zeller <>
Cc: "" <>
Subject: Re: [grouper-users] PSP problem after folder rename

Thanks, guys, it's working now.  Tom, yes, when I changed the sequence number, I did double-check the table to be sure it was right, however when I checked again this morning, it hadn't really made it to the db.  After I got that change in for real, it took off.  BTW, what's the significance of hibernate_version_number?

Tim Darby
The University of Arizona
Mosaic, Systems Integration and Architecture

UITS, Rm 335, 520-626-3799


On Fri, May 17, 2013 at 10:37 AM, Tom Zeller <> wrote:
I thought each consumer stores its last sequence number somewhere, so
my initial response is are you sure you changed the psp's ?

On Fri, May 17, 2013 at 11:07 AM, Tim Darby <> wrote:
> I had a similar thing with change log processing, so I tried advancing the
> psp sequence number in grouper_change_log_consumer.  However, when I restart
> gsh -loader, psp starts with a sequence number prior to the one I set, tries
> 100 sequence numbers past that and stops.  It now won't go past that point
> and acts like it's broken.  The other two consumers, syncGroups and
> grouperRules, are still being updated correctly, but psp is dead in the
> water.  Why is it looking at sequence numbers smaller than the one in
> grouper_change_log_consumer and what can I do to get it going again?
>
> Tim Darby
> The University of Arizona
> Mosaic, Systems Integration and Architecture
>
> UITS, Rm 335, 520-626-3799
>
>
> On Wed, May 15, 2013 at 7:08 AM, David Langenberg <>
> wrote:
>>
>> Hi Bryan,
>>
>> The table grouper_change_log_consumer holds the pointer.
>>
>> Dave
>>
>>
>> On Wed, May 15, 2013 at 8:04 AM, Bryan E. Wooten <>
>> wrote:
>>>
>>> Dave,
>>>
>>> Thanks for the explanation, that makes sense and adds to my understand of
>>> how the change log works.
>>>
>>> Can you point me to the table I need to modify? (i.e. Where is the change
>>> log sequence number?)
>>>
>>> Thanks,
>>>
>>> Bryan
>>>
>>> From: David Langenberg <>
>>> Date: Tuesday, May 14, 2013 11:37 PM
>>> To: Bryan Wooten <>
>>> Cc: "" <>
>>> Subject: Re: [grouper-users] PSP problem after folder rename
>>>
>>> The PSP is trying to process this entry because BulkSync doesn't advance
>>> the PSP change log consumer to the most recent event.  Think of it this way,
>>> the PSP Change log consumer is pointing at change log sequence 100.  You
>>> then rename the stem, a bunch of change log entries are added, however,
>>> because the change log consumer isn't actively processing them the pointer
>>> is still at 100.  You then run the BulkSync.  BulkSync does an immediate
>>> calc/diff of grouper to your AD and performs the rename operation.  You then
>>> tell the change log processor to start processing.  It is still at sequence
>>> 100 and advances to the rename operation, let's call this sequence 105.  105
>>> says rename from A to B.  However, BulkSync already performed that
>>> operation.  So when the consumer goes looking for A in AD, it doesn't find
>>> it and you get this error.  What should probably happen in this instance is
>>> the PSP attempt to see if the rename operation already happened (possibly
>>> schedule a full-calc of B to ensure that it did get correctly re-named
>>> instead of deleted by some other process).
>>>
>>> For now though, I recommend you advance the change log pointer for the
>>> PSP past this problem entry and let it continue processing.
>>>
>>> Dave
>>>
>>>
>>> On Tue, May 14, 2013 at 11:34 AM, Bryan E. Wooten <>
>>> wrote:
>>>>
>>>> After renaming a folder and getting ./gsh  -psp –bulkSync to complete
>>>> without errors my AD now reflects my Grouper folder and group structure
>>>> accurately.
>>>>
>>>>
>>>>
>>>> I added a folder using the UI and attempted to run:
>>>> loaderRunOneJob("CHANGE_LOG_changeLogTempToChangeLog")
>>>>
>>>> and loaderRunOneJob("CHANGE_LOG_consumer_psp").
>>>>
>>>>
>>>>
>>>> But I get the following error:
>>>>
>>>>
>>>>
>>>> - An error occurred processing sequence number 192889
>>>>
>>>> edu.internet2.middleware.psp.PspException:
>>>> edu.internet2.middleware.psp.PspNoSuchIdentifierException: Psp 'psp' - Has
>>>> reference from
>>>> 'PSOIdentifier[id='CN=uuCurrentEmployee,OU=groups,OU=TestStem,OU=uofu,OU=groups,OU=grouper,DC=testad,DC=utah,DC=edu',targetID=ldap,containerID=<null>]'
>>>> to
>>>> 'Reference[toPsoID=PSOIdentifier[id='CN=u0530608,ou=people,dc=testad,dc=utah,dc=edu',targetID=ldap,containerID=<null>],type=member]'
>>>> SearchResponse[psos=0,status=failure,error=noSuchIdentifier,errorMessages={[LDAP:
>>>> error code 32 - 0000208D: NameErr: DSID-0310020A, problem 2001 (NO_OBJECT),
>>>> data 0, best match of:
>>>>
>>>>         'OU=uofu,OU=groups,OU=grouper,DC=testad,DC=utah,DC=edu'
>>>>
>>>>
>>>>
>>>> The OU TestStem is the folder we renamed and the bulk sync removed the
>>>> OU from AD as expected.
>>>>
>>>>
>>>>
>>>> It is unclear why there is still an entry for this folder in the Change
>>>> Log. When this error occurs the Change Log Consumer stops processing so my
>>>> new created folder does not get provisioned to AD.
>>>>
>>>>
>>>>
>>>> Is this just a rogue entry in the Change Log? I am sure this is not the
>>>> only one.  Should I just remove all the entries in the change log to fix
>>>> this issue or is there a better solution?
>>>>
>>>>
>>>>
>>>> Thanks,
>>>>
>>>>
>>>>
>>>> Bryan
>>>
>>>
>>>
>>>
>>> --
>>> David Langenberg
>>> Identity & Access Management
>>> The University of Chicago
>>
>>
>>
>>
>> --
>> David Langenberg
>> Identity & Access Management
>> The University of Chicago
>
>




Archive powered by MHonArc 2.6.16.

Top of Page