Grouper Developers Forum

["Gettes, Michael" <>]

I got a 3 runs of findBadMemberships to be 0 bad memberships in the first and 
only pass.

Re-ran the loader and the 14 or so groups remain an error.

/mrg

> On May 31, 2019, at 4:49 PM, Gettes, Michael <> wrote:
> 
> and it does iterate a few times.  I don’t think this is going to fix these 
> problems.
> 
> can you offer advice on what i should be deleting out of grouper_group_set? 
>  I am pretty fBM is not going to fix this issue.
> 
> /mrg
> 
>> On May 31, 2019, at 4:36 PM, Gettes, Michael <> wrote:
>> 
>> I run findBadMemberships as Other_JOB.  I will keep running it until it 
>> says no bad memberships.  Maybe that will fix it?
>> 
>> /mrg
>> 
>>> On May 31, 2019, at 4:21 PM, Shilen Patel <> wrote:
>>> 
>>> I wonder why it’s finding so many issues.  If you run it again, does it 
>>> find more?  You can run it via the daemon which would keep running it 
>>> until it doesn’t find anymore (or quit after 4 attempts).  Before going 
>>> into the db and deleting rows out of grouper_group_set, it would be good 
>>> to know if this fixes it or not.
>>> 
>>> Thanks!
>>> 
>>> - Shilen
>>> 
>>> 
>>> On 5/31/19, 4:03 PM, "Gettes, Michael" <> wrote:
>>> 
>>>  Shilen - findBadMemberships found 3 bad or missing this time around.
>>> 
>>>  Re-ran the loader job again and i still get the 14-ish errored groups as 
>>> before.
>>> 
>>>  /mrg
>>> 
>>>> On May 31, 2019, at 2:49 PM, Gettes, Michael <> wrote:
>>>> 
>>>> at 2:56am this morning findBadMemberships reported:
>>>> 
>>>> Found 8 bad or missing memberships. Fixing and will try again. Found 1 
>>>> bad or missing memberships. Fixing and will try again. Found 42 bad or 
>>>> missing memberships. Fixing and will try again. Found 0 bad or missing 
>>>> memberships. 
>>>> 
>>>> running it again.
>>>> 
>>>> /mrg
>>>> 
>>>>> On May 31, 2019, at 2:47 PM, Gettes, Michael <> wrote:
>>>>> 
>>>>> my last response was before I read this.  running findBadMemberships 
>>>>> now which runs every nite at 2am.
>>>>> 
>>>>> Carey - i hope my previous response sorta answers your questions.
>>>>> 
>>>>> /mrg
>>>>> 
>>>>>> On May 31, 2019, at 2:43 PM, Shilen Patel <> wrote:
>>>>>> 
>>>>>> Have you tried to run findBadMemberships()?  Depending on what the 
>>>>>> group set issue is, that may or may not help.  
>>>>>> 
>>>>>> - Shilen
>>>>>> 
>>>>>> On 5/31/19, 1:43 PM, " on behalf of 
>>>>>> Gettes, Michael" < on behalf of 
>>>>>> > wrote:
>>>>>> 
>>>>>> I’m trying to decipher the error presented…
>>>>>> 
>>>>>> Caused by: java.sql.BatchUpdateException: Cannot delete or update a 
>>>>>> parent row: a foreign key constraint fails 
>>>>>> (`grouper`.`grouper_group_set`, CONSTRAINT 
>>>>>> `fk_group_set_member_group_id` FOREIGN KEY (`member_group_id`) 
>>>>>> REFERENCES `grouper_groups` (`id`))
>>>>>> 
>>>>>> I read this to imply there is a value in 
>>>>>> grouper_group_set.member_group_id that is not in grouper_groups.id.
>>>>>> To prove, I ran a right outer join to find what’s in group_set that is 
>>>>>> not in groups
>>>>>> select * from grouper.grouper_group_set gs 
>>>>>> right join grouper.grouper_groups g 
>>>>>> on gs.member_group_id = g.id
>>>>>> where gs.member_group_id = null
>>>>>> 
>>>>>> and got 0 records returned.
>>>>>> 
>>>>>> i am not sure where the problem is - any guidance?
>>>>>> 
>>>>>>> On May 31, 2019, at 11:16 AM, Gettes, Michael <> wrote:
>>>>>>> 
>>>>>>> Hi all,
>>>>>>> 
>>>>>>> I have a condition of about a dozen groups which cannot be modified 
>>>>>>> or deleted and the log snippet attached shows the error log for one 
>>>>>>> of these groups.  First, I’d like to know how I fix these groups.  
>>>>>>> Second, the condition of how this occurred I am still trying to prove 
>>>>>>> but it seems to be related to whether or not memory is configured 
>>>>>>> with ms=mx.  I haven’t been able to recreate what caused the 
>>>>>>> condition in the log snippet but the log snippet with ms=mx devolves 
>>>>>>> into Garbarge collection failures and the loader becoming unusable 
>>>>>>> over time.  With ms<mx (8g<9g) the errors i see in the log snippet do 
>>>>>>> not cause the loader to become unusable over time.  Anyway, I’d 
>>>>>>> really appreciate help fixing the bad groups based on the log.
>>>>>>> 
>>>>>>> I decided not to post this to the grouper-users list because I don’t 
>>>>>>> want people thinking grouper is unstable and this could happen to 
>>>>>>> them.  As you guys know, i have been doing some extreme things with 
>>>>>>> grouper to see what’s possible and find issues and i think these 
>>>>>>> “broken groups” are related to that work and the various memory 
>>>>>>> problems I was having.
>>>>>>> 
>>>>>>> Thank you!
>>>>>>> 
>>>>>>> /mrg
>>>>>>> 
>>>>>>> <20190531-grouper-error.log>
>>>>>> 
>>>>>> 
>>>>>> 
>>>>> 
>>>> 
>>> 
>>> 
>>> 
>> 
>