You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@solr.apache.org by dmitri maziuk <dm...@gmail.com> on 2023/02/17 00:48:35 UTC

8.11.2 missing record after adding child docs

Hi all,

I'm having a FTW moment with another install of 8.11.2. I am posting a 
parent document with "TaxCreditId" field = 1 and a couple of child 
documents. Parent's id is generated as uuid5, i.e. it's stable. Core 
name is WHPD.

When I post the parent only here's what I get from querying 
WHPD/select?indent=3Dtrue&q.op=3DOR&q=3DTax=
CreditId%3A* :
'''
{
   "responseHeader":{
     "status":0,
     "QTime":0,
     "params":{
       "q":"TaxCreditId:*",
       "indent":"true",
       "q.op":"OR",
       "_":"1676591103290"}},
   "response":{"numFound":1,"start":0,"numFoundExact":true,"docs":[
       {
         "TaxCreditId":1,
         "id":"3ff1833a-ea58-5707-8bb3-77add16d08f2",
...
'''

When I post the children as well, here's what I find:
'''
{
   "responseHeader":{
     "status":0,
     "QTime":0,
     "params":{
       "q":"TaxCreditId:*",
       "indent":"true",
       "q.op":"OR",
       "_":"1676591103290"}},
   "response":{"numFound":0,"start":0,"numFoundExact":true,"docs":[]
   }}
'''

The log says it all got posted:
'''
2023-02-17 00:29:38.766 INFO  (qtp178604517-1260) [   x:WHPD] 
o.a.s.u.p.Log=UpdateProcessorFactory [WHPD]  webapp=3D/solr 
path=3D/update 
params=3D{comm=it=3Dtrue}{add=3D[3ff1833a-ea58-5707-8bb3-77add16d08f2 
(1758035998122442752=)],commit=3D} 0 27
2023-02-17 00:29:38.800 INFO 
(searcherExecutor-33-thread-1-processing-x:WH=PD) [   x:WHPD] 
o.a.s.c.QuerySenderListener QuerySenderListener done.
2023-02-17 00:29:38.801 INFO 
(searcherExecutor-33-thread-1-processing-x:WH=PD) [   x:WHPD] 
o.a.s.c.SolrCore [WHPD]  Registered new searcher autowarm t=ime: 0 ms
2023-02-17 00:29:38.803 INFO  (qtp178604517-1265) [   x:WHPD] 
o.a.s.u.p.Log=UpdateProcessorFactory [WHPD]  webapp=3D/solr 
path=3D/update params=3D{comm=it=3Dtrue}{add=3D[TCLetter1 
(root=3D3ff1833a-ea58-5707-8bb3-77add16d08f2) (=1758035998174871552), 
TCLetter2 (root=3D3ff1833a-ea58-5707-8bb3-77add16d08f=2) 
(1758035998174871553), TCLetter12 
(root=3D3ff1833a-ea58-5707-8bb3-77add1=6d08f2) (1758035998175920128), 
TCLetter1223 (root=3D3ff1833a-ea58-5707-8bb3=-77add16d08f2) 
(1758035998175920129), TCWorkItem1 
(root=3D3ff1833a-ea58-570=7-8bb3-77add16d08f2) (1758035998175920130), 
TCWorkItem2 (root=3D3ff1833a-ea=58-5707-8bb3-77add16d08f2) 
(1758035998175920131), TCWorkItem3 
(root=3D3ff18=33a-ea58-5707-8bb3-77add16d08f2) 
(1758035998175920132)],commit=3D} 0 26
'''

Any idea what could be causing this? Or where to even start looking?

TIA
Dima


Re: 8.11.2 missing record after adding child docs -- BUG?

Posted by Mikhail Khludnev <mk...@apache.org>.
Well.. I missed the fact that implementation went so far already.
This particular case
https://solr.apache.org/guide/8_11/updating-parts-of-documents.html#updating-child-documents
should work, I believe.


On Fri, Feb 17, 2023 at 11:27 PM dmitri maziuk <dm...@gmail.com>
wrote:

> On 2023-02-17 2:14 PM, Mikhail Khludnev wrote:
> > Pardon. I'm lost.
> > Are you using
> >
> https://solr.apache.org/guide/solr/latest/indexing-guide/indexing-nested-documents.html
> > ?
> > With that approach it's not possible to put parents and children
> > separately, they go strictly altogether.
>
> So this:
> https://solr.apache.org/guide/8_11/updating-parts-of-documents.html is
> not true? And the fact that it works on one of two servers is a fluke?
>
> Dima
>
>

-- 
Sincerely yours
Mikhail Khludnev
https://t.me/MUST_SEARCH
A caveat: Cyrillic!

Re: 8.11.2 missing record after adding child docs -- BUG?

Posted by dmitri maziuk <dm...@gmail.com>.
On 2023-02-17 2:14 PM, Mikhail Khludnev wrote:
> Pardon. I'm lost.
> Are you using
> https://solr.apache.org/guide/solr/latest/indexing-guide/indexing-nested-documents.html
> ?
> With that approach it's not possible to put parents and children
> separately, they go strictly altogether.

So this: 
https://solr.apache.org/guide/8_11/updating-parts-of-documents.html is 
not true? And the fact that it works on one of two servers is a fluke?

Dima


Re: 8.11.2 missing record after adding child docs -- FIXED

Posted by dmitri maziuk <dm...@gmail.com>.
OK, I am officially an idiot. Apologies for the noise everyone.

(The answer is there was one uncommitted change in the importer script 
hiding on the development host -- something `git pull origin` doesn't 
tell you about. Bad git.)

Cheers,
Dima


Re: 8.11.2 missing record after adding child docs -- BUG?

Posted by Mikhail Khludnev <mk...@apache.org>.
Pardon. I'm lost.
Are you using
https://solr.apache.org/guide/solr/latest/indexing-guide/indexing-nested-documents.html
?
With that approach it's not possible to put parents and children
separately, they go strictly altogether.

On Fri, Feb 17, 2023 at 8:22 PM dmitri maziuk <dm...@gmail.com>
wrote:

> On 2023-02-16 11:59 PM, Mikhail Khludnev wrote:
> > Hello,
> > May all of them clash by uniqueKey? If it has place children may wipe
> > parent. What do you get on *:*?
>
> uniqueKey does not clash between parent and child documents. Importer
> POSTs all parents first, with "type" : "parent" and "id" : uuid5 derived
> from (parent's) DB id.
>
> Then it POSTs all children with "id" : (child's) DB id and "_root_" :
> uuid5 derived from parent's DB id.
>
> It is actually working on another Solr node whose setup should be
> identical -- at least I don't see any differences -- so I know there is
> nothing wrong with what I am inserting.
>
> So is there a way to find out why exactly it's wiping out the parents?
> Or is it just a heisenbug in the twisted maze of little update
> processors, all alike?
>
> Dima
>
>

-- 
Sincerely yours
Mikhail Khludnev
https://t.me/MUST_SEARCH
A caveat: Cyrillic!

Re: 8.11.2 missing record after adding child docs -- BUG?

Posted by dmitri maziuk <dm...@gmail.com>.
On 2023-02-16 11:59 PM, Mikhail Khludnev wrote:
> Hello,
> May all of them clash by uniqueKey? If it has place children may wipe
> parent. What do you get on *:*?

uniqueKey does not clash between parent and child documents. Importer 
POSTs all parents first, with "type" : "parent" and "id" : uuid5 derived 
from (parent's) DB id.

Then it POSTs all children with "id" : (child's) DB id and "_root_" : 
uuid5 derived from parent's DB id.

It is actually working on another Solr node whose setup should be 
identical -- at least I don't see any differences -- so I know there is 
nothing wrong with what I am inserting.

So is there a way to find out why exactly it's wiping out the parents? 
Or is it just a heisenbug in the twisted maze of little update 
processors, all alike?

Dima


Re: 8.11.2 missing record after adding child docs

Posted by Mikhail Khludnev <mk...@apache.org>.
Hello,
May all of them clash by uniqueKey? If it has place children may wipe
parent. What do you get on *:*?
FWIW, logs excerpt is grabbed with 3D something, not really readable.

On Fri, Feb 17, 2023 at 3:48 AM dmitri maziuk <dm...@gmail.com>
wrote:

> Hi all,
>
> I'm having a FTW moment with another install of 8.11.2. I am posting a
> parent document with "TaxCreditId" field = 1 and a couple of child
> documents. Parent's id is generated as uuid5, i.e. it's stable. Core
> name is WHPD.
>
> When I post the parent only here's what I get from querying
> WHPD/select?indent=3Dtrue&q.op=3DOR&q=3DTax=
> CreditId%3A* :
> '''
> {
>    "responseHeader":{
>      "status":0,
>      "QTime":0,
>      "params":{
>        "q":"TaxCreditId:*",
>        "indent":"true",
>        "q.op":"OR",
>        "_":"1676591103290"}},
>    "response":{"numFound":1,"start":0,"numFoundExact":true,"docs":[
>        {
>          "TaxCreditId":1,
>          "id":"3ff1833a-ea58-5707-8bb3-77add16d08f2",
> ...
> '''
>
> When I post the children as well, here's what I find:
> '''
> {
>    "responseHeader":{
>      "status":0,
>      "QTime":0,
>      "params":{
>        "q":"TaxCreditId:*",
>        "indent":"true",
>        "q.op":"OR",
>        "_":"1676591103290"}},
>    "response":{"numFound":0,"start":0,"numFoundExact":true,"docs":[]
>    }}
> '''
>
> The log says it all got posted:
> '''
> 2023-02-17 00:29:38.766 INFO  (qtp178604517-1260) [   x:WHPD]
> o.a.s.u.p.Log=UpdateProcessorFactory [WHPD]  webapp=3D/solr
> path=3D/update
> params=3D{comm=it=3Dtrue}{add=3D[3ff1833a-ea58-5707-8bb3-77add16d08f2
> (1758035998122442752=)],commit=3D} 0 27
> 2023-02-17 00:29:38.800 INFO
> (searcherExecutor-33-thread-1-processing-x:WH=PD) [   x:WHPD]
> o.a.s.c.QuerySenderListener QuerySenderListener done.
> 2023-02-17 00:29:38.801 INFO
> (searcherExecutor-33-thread-1-processing-x:WH=PD) [   x:WHPD]
> o.a.s.c.SolrCore [WHPD]  Registered new searcher autowarm t=ime: 0 ms
> 2023-02-17 00:29:38.803 INFO  (qtp178604517-1265) [   x:WHPD]
> o.a.s.u.p.Log=UpdateProcessorFactory [WHPD]  webapp=3D/solr
> path=3D/update params=3D{comm=it=3Dtrue}{add=3D[TCLetter1
> (root=3D3ff1833a-ea58-5707-8bb3-77add16d08f2) (=1758035998174871552),
> TCLetter2 (root=3D3ff1833a-ea58-5707-8bb3-77add16d08f=2)
> (1758035998174871553), TCLetter12
> (root=3D3ff1833a-ea58-5707-8bb3-77add1=6d08f2) (1758035998175920128),
> TCLetter1223 (root=3D3ff1833a-ea58-5707-8bb3=-77add16d08f2)
> (1758035998175920129), TCWorkItem1
> (root=3D3ff1833a-ea58-570=7-8bb3-77add16d08f2) (1758035998175920130),
> TCWorkItem2 (root=3D3ff1833a-ea=58-5707-8bb3-77add16d08f2)
> (1758035998175920131), TCWorkItem3
> (root=3D3ff18=33a-ea58-5707-8bb3-77add16d08f2)
> (1758035998175920132)],commit=3D} 0 26
> '''
>
> Any idea what could be causing this? Or where to even start looking?
>
> TIA
> Dima
>
>

-- 
Sincerely yours
Mikhail Khludnev
https://t.me/MUST_SEARCH
A caveat: Cyrillic!