You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@hawq.apache.org by Brian Willingham <br...@ericsson.com> on 2017/01/31 13:57:23 UTC
UNSUBSCRIBE
UNSUBSCRIBE
From: Alex (Oleksandr) Diachenko [mailto:odiachenko@pivotal.io]
Sent: Monday, January 30, 2017 1:52 PM
To: dev@hawq.incubator.apache.org
Cc: user@hawq.incubator.apache.org
Subject: Re: PXF HCatalog Issue for Non-Super User
Gagan,
Correct, that's was a mandatory condition for HCatalog intergration to work properly before the HAWQ-1130 fix.
Regards, Alex.
On Mon, Jan 30, 2017 at 11:30 AM, Gagan Brahmi <ga...@gmail.com>> wrote:
Appreciate that info Alex. Granting SELECT or SUPERUSER only viable
workaround for this problem for now. Is that a right understanding?
Regards,
Gagan Brahmi
On Mon, Jan 30, 2017 at 12:07 PM, Alex (Oleksandr) Diachenko <
odiachenko@pivotal.io<ma...@pivotal.io>> wrote:
> Hi Gagan,
>
> The issue you described was a real limitation before
> https://issues.apache.org/jira/browse/HAWQ-1130 was reported.
> As for now the issue was resolved and targeted for next 2.1.0.0-incubating
> open-source release.
>
> Regards, Alex.
>
>
> On Mon, Jan 30, 2017 at 10:08 AM, Gagan Brahmi <ga...@gmail.com>>
> wrote:
>
>> Hi All,
>>
>> I am not sure if anyone has faced issue for a non-privileged HAWQ user
>> while access hive table using HCatalog. The issue is encountered when a
>> non-privileged user tries to query the hive table (hive_table) under any
>> database (default in this example) using hcatalog
>> (hcatalog.default.hive_table).
>>
>> However, when you describe the table (\d) there is no issue encountered.
>>
>> I have tested this on HDB 2.0.0 and HDB 2.0.1.
>>
>> For the non-privileged user the following is the behavior when trying to
>> access the data using SELECT:
>>
>> --------------------
>>
>> postgres=> SELECT * FROM hcatalog.default.hive_table;
>> ERROR: permission denied for relation pg_authid
>> LINE 1: SELECT COUNT(*) FROM hcatalog.default.hive_table;
>> ^
>> CONTEXT: SQL statement "SELECT max(oid) FROM (SELECT max(oid) AS oid FROM
>> pg_extprotocol UNION ALL SELECT max(oid) AS oid FROM pg_partition UNION
>> ALL
>> SELECT max(oid) AS oid FROM pg_partition_rule UNION ALL SELECT max(oid) AS
>> oid FROM pg_filespace UNION ALL SELECT max(oid) AS oid FROM pg_compression
>> UNION ALL SELECT max(oid) AS oid FROM pg_filesystem UNION ALL SELECT
>> max(oid) AS oid FROM pg_foreign_data_wrapper UNION ALL SELECT max(oid) AS
>> oid FROM pg_foreign_server UNION ALL SELECT max(oid) AS oid FROM
>> pg_database UNION ALL SELECT max(oid) AS oid FROM pg_type UNION ALL SELECT
>> max(oid) AS oid FROM pg_proc UNION ALL SELECT max(oid) AS oid FROM
>> pg_class
>> UNION ALL SELECT max(oid) AS oid FROM pg_attrdef UNION ALL SELECT max(oid)
>> AS oid FROM pg_constraint UNION ALL SELECT max(oid) AS oid FROM
>> pg_operator
>> UNION ALL SELECT max(oid) AS oid FROM pg_opclass UNION ALL SELECT max(oid)
>> AS oid FROM pg_am UNION ALL SELECT max(oid) AS oid FROM pg_language UNION
>> ALL SELECT max(oid) AS oid FROM pg_rewrite UNION ALL SELECT max(oid) AS
>> oid
>> FROM pg_trigger UNION ALL SELECT max(oid) AS oid FROM pg_cast UNION ALL
>> SELECT max(oid) AS oid FROM pg_namespace UNION ALL SELECT max(oid) AS oid
>> FROM pg_conversion UNION ALL SELECT max(oid) AS oid FROM pg_tablespace
>> UNION ALL SELECT max(oid) AS oid FROM pg_resqueue UNION ALL SELECT
>> max(oid)
>> AS oid FROM pg_authid UNION ALL SELECT max(oid) AS oid FROM
>> pg_user_mapping) AS x"
>>
>> postgres=> SELECT * FROM hcatalog.default.hive_table;
>> ERROR: database does not have enough available Oids to support HCatalog
>> queries
>> LINE 1: SELECT * FROM hcatalog.default.hive_table;
>> ^
>> HINT: Database VACUUM may recycle unused Oids.
>>
>>
>> postgres=> \d hcatalog.default.hive_table
>>
>> PXF Hive Table
>> "default.hive_table"
>> Column | Type
>> --------+------
>> id | int4
>> fname | text
>> lname | text
>>
>> --------------------
>>
>> I was able to get around this problem by granting SELECT on pg_authid and
>> pg_user_mapping table which fixes the underlying query to grab the max
>> Oid.
>>
>> Is this is a known issue or is there any better workaround for this
>> behavior? Or anything other than granting the user SUPERUSER privileges?
>>
>>
>> Regards,
>> Gagan Brahmi
>>
>
>
Re: UNSUBSCRIBE
Posted by Marc Paquette <mp...@pivotal.io>.
Hi Folks,
To unsubscribe please email *user-unsubscribe@hawq.incubator.apache.org
<us...@hawq.incubator.apache.org>*
Thank you,
Marc Paquette
On Jan 31, 2017 at 09:21 AM, "swethak yadav" <sw...@gmail.com>
wrote:
> Unsubscribe
>
> On Jan 31, 2017 9:16 AM, "Russ Holmes" <rh...@hortonworks.com> wrote:
>
>> Unsubscribe.
>>
>>
>> ----------------------------------------------------------------------
>>
>> *Russell Holmes*
>>
>> Hortonworks Public Sector
>>
>> +1.703.728.0669 <(703)%20728-0669>
>>
>> <https://twitter.com/zimandan>
>>
>> Hadoop/DataWorks Summit <https://dataworkssummit.com/san-jose-2017/>
>> San Jose, CA | June 13-15
>> ------------------------------
>> *From:* Supriya Biswas <su...@nielsen.com>
>> *Sent:* Tuesday, January 31, 2017 9:03 AM
>> *To:* Contact
>> *Subject:* Re: UNSUBSCRIBE
>>
>> UNSUBSCRIBE
>>
>>
>> *Supriyo Biswas *Architect
>> The Nielsen Company
>> Office (516) 682-6021/NETS 249-6021
>>
>> Cell (516) 353-6795
>> www.nielsen.com
>>
>> On Tue, Jan 31, 2017 at 8:57 AM, Brian Willingham <
>> brian.willingham@ericsson.com> wrote:
>>
>>> UNSUBSCRIBE
>>>
>>>
>>>
>>>
>>>
>>> *From:* Alex (Oleksandr) Diachenko [mailto:odiachenko@pivotal.io]
>>> *Sent:* Monday, January 30, 2017 1:52 PM
>>> *To:* dev@hawq.incubator.apache.org
>>> *Cc:* user@hawq.incubator.apache.org
>>> *Subject:* Re: PXF HCatalog Issue for Non-Super User
>>>
>>>
>>>
>>> Gagan,
>>>
>>>
>>>
>>> Correct, that's was a mandatory condition for HCatalog intergration to
>>> work properly before the HAWQ-1130 fix.
>>>
>>>
>>>
>>> Regards, Alex.
>>>
>>>
>>>
>>> On Mon, Jan 30, 2017 at 11:30 AM, Gagan Brahmi <ga...@gmail.com>
>>> wrote:
>>>
>>> Appreciate that info Alex. Granting SELECT or SUPERUSER only viable
>>> workaround for this problem for now. Is that a right understanding?
>>>
>>>
>>> Regards,
>>> Gagan Brahmi
>>>
>>>
>>> On Mon, Jan 30, 2017 at 12:07 PM, Alex (Oleksandr) Diachenko <
>>> odiachenko@pivotal.io> wrote:
>>>
>>> > Hi Gagan,
>>> >
>>> > The issue you described was a real limitation before
>>> > https://issues.apache.org/jira/browse/HAWQ-1130 was reported.
>>> > As for now the issue was resolved and targeted for next
>>> 2.1.0.0-incubating
>>> > open-source release.
>>> >
>>> > Regards, Alex.
>>> >
>>> >
>>> > On Mon, Jan 30, 2017 at 10:08 AM, Gagan Brahmi <ga...@gmail.com>
>>> > wrote:
>>> >
>>> >> Hi All,
>>> >>
>>> >> I am not sure if anyone has faced issue for a non-privileged HAWQ user
>>> >> while access hive table using HCatalog. The issue is encountered when
>>> a
>>> >> non-privileged user tries to query the hive table (hive_table) under
>>> any
>>> >> database (default in this example) using hcatalog
>>> >> (hcatalog.default.hive_table).
>>> >>
>>> >> However, when you describe the table (\d) there is no issue
>>> encountered.
>>> >>
>>> >> I have tested this on HDB 2.0.0 and HDB 2.0.1.
>>> >>
>>> >> For the non-privileged user the following is the behavior when trying
>>> to
>>> >> access the data using SELECT:
>>> >>
>>> >> --------------------
>>> >>
>>> >> postgres=> SELECT * FROM hcatalog.default.hive_table;
>>> >> ERROR: permission denied for relation pg_authid
>>> >> LINE 1: SELECT COUNT(*) FROM hcatalog.default.hive_table;
>>> >> ^
>>> >> CONTEXT: SQL statement "SELECT max(oid) FROM (SELECT max(oid) AS oid
>>> FROM
>>> >> pg_extprotocol UNION ALL SELECT max(oid) AS oid FROM pg_partition
>>> UNION
>>> >> ALL
>>> >> SELECT max(oid) AS oid FROM pg_partition_rule UNION ALL SELECT
>>> max(oid) AS
>>> >> oid FROM pg_filespace UNION ALL SELECT max(oid) AS oid FROM
>>> pg_compression
>>> >> UNION ALL SELECT max(oid) AS oid FROM pg_filesystem UNION ALL SELECT
>>> >> max(oid) AS oid FROM pg_foreign_data_wrapper UNION ALL SELECT
>>> max(oid) AS
>>> >> oid FROM pg_foreign_server UNION ALL SELECT max(oid) AS oid FROM
>>> >> pg_database UNION ALL SELECT max(oid) AS oid FROM pg_type UNION ALL
>>> SELECT
>>> >> max(oid) AS oid FROM pg_proc UNION ALL SELECT max(oid) AS oid FROM
>>> >> pg_class
>>> >> UNION ALL SELECT max(oid) AS oid FROM pg_attrdef UNION ALL SELECT
>>> max(oid)
>>> >> AS oid FROM pg_constraint UNION ALL SELECT max(oid) AS oid FROM
>>> >> pg_operator
>>> >> UNION ALL SELECT max(oid) AS oid FROM pg_opclass UNION ALL SELECT
>>> max(oid)
>>> >> AS oid FROM pg_am UNION ALL SELECT max(oid) AS oid FROM pg_language
>>> UNION
>>> >> ALL SELECT max(oid) AS oid FROM pg_rewrite UNION ALL SELECT max(oid)
>>> AS
>>> >> oid
>>> >> FROM pg_trigger UNION ALL SELECT max(oid) AS oid FROM pg_cast UNION
>>> ALL
>>> >> SELECT max(oid) AS oid FROM pg_namespace UNION ALL SELECT max(oid) AS
>>> oid
>>> >> FROM pg_conversion UNION ALL SELECT max(oid) AS oid FROM pg_tablespace
>>> >> UNION ALL SELECT max(oid) AS oid FROM pg_resqueue UNION ALL SELECT
>>> >> max(oid)
>>> >> AS oid FROM pg_authid UNION ALL SELECT max(oid) AS oid FROM
>>> >> pg_user_mapping) AS x"
>>> >>
>>> >> postgres=> SELECT * FROM hcatalog.default.hive_table;
>>> >> ERROR: database does not have enough available Oids to support
>>> HCatalog
>>> >> queries
>>> >> LINE 1: SELECT * FROM hcatalog.default.hive_table;
>>> >> ^
>>> >> HINT: Database VACUUM may recycle unused Oids.
>>> >>
>>> >>
>>> >> postgres=> \d hcatalog.default.hive_table
>>> >>
>>> >> PXF Hive Table
>>> >> "default.hive_table"
>>> >> Column | Type
>>> >> --------+------
>>> >> id | int4
>>> >> fname | text
>>> >> lname | text
>>> >>
>>> >> --------------------
>>> >>
>>> >> I was able to get around this problem by granting SELECT on pg_authid
>>> and
>>> >> pg_user_mapping table which fixes the underlying query to grab the max
>>> >> Oid.
>>> >>
>>> >> Is this is a known issue or is there any better workaround for this
>>> >> behavior? Or anything other than granting the user SUPERUSER
>>> privileges?
>>> >>
>>> >>
>>> >> Regards,
>>> >> Gagan Brahmi
>>> >>
>>> >
>>> >
>>>
>>>
>>>
>>
>>
--
Marc Paquette | Designated Support Engineer | 215-346-6272
Support.Pivotal.io <http://support.pivotal.io/> | Mon-Fri 9:30am to
5:00pm EST | 1-877-477-2269
[image: support] <https://support.pivotal.io/> [image: twitter]
<https://twitter.com/pivotal> [image: linkedin]
<https://www.linkedin.com/company/3048967> [image: facebook]
<https://www.facebook.com/pivotalsoftware> [image: google plus]
<https://plus.google.com/+Pivotal> [image: youtube]
<https://www.youtube.com/playlist?list=PLAdzTan_eSPScpj2J50ErtzR9ANSzv3kl>
Re: UNSUBSCRIBE
Posted by swethak yadav <sw...@gmail.com>.
Unsubscribe
On Jan 31, 2017 9:16 AM, "Russ Holmes" <rh...@hortonworks.com> wrote:
> Unsubscribe.
>
>
> ----------------------------------------------------------------------
>
> *Russell Holmes*
>
> Hortonworks Public Sector
>
> +1.703.728.0669 <(703)%20728-0669>
>
> <https://twitter.com/zimandan>
>
> Hadoop/DataWorks Summit <https://dataworkssummit.com/san-jose-2017/>
> San Jose, CA | June 13-15
> ------------------------------
> *From:* Supriya Biswas <su...@nielsen.com>
> *Sent:* Tuesday, January 31, 2017 9:03 AM
> *To:* Contact
> *Subject:* Re: UNSUBSCRIBE
>
> UNSUBSCRIBE
>
>
> *Supriyo Biswas *Architect
> The Nielsen Company
> Office (516) 682-6021/NETS 249-6021
>
> Cell (516) 353-6795
> www.nielsen.com
>
> On Tue, Jan 31, 2017 at 8:57 AM, Brian Willingham <
> brian.willingham@ericsson.com> wrote:
>
>> UNSUBSCRIBE
>>
>>
>>
>>
>>
>> *From:* Alex (Oleksandr) Diachenko [mailto:odiachenko@pivotal.io]
>> *Sent:* Monday, January 30, 2017 1:52 PM
>> *To:* dev@hawq.incubator.apache.org
>> *Cc:* user@hawq.incubator.apache.org
>> *Subject:* Re: PXF HCatalog Issue for Non-Super User
>>
>>
>>
>> Gagan,
>>
>>
>>
>> Correct, that's was a mandatory condition for HCatalog intergration to
>> work properly before the HAWQ-1130 fix.
>>
>>
>>
>> Regards, Alex.
>>
>>
>>
>> On Mon, Jan 30, 2017 at 11:30 AM, Gagan Brahmi <ga...@gmail.com>
>> wrote:
>>
>> Appreciate that info Alex. Granting SELECT or SUPERUSER only viable
>> workaround for this problem for now. Is that a right understanding?
>>
>>
>> Regards,
>> Gagan Brahmi
>>
>>
>> On Mon, Jan 30, 2017 at 12:07 PM, Alex (Oleksandr) Diachenko <
>> odiachenko@pivotal.io> wrote:
>>
>> > Hi Gagan,
>> >
>> > The issue you described was a real limitation before
>> > https://issues.apache.org/jira/browse/HAWQ-1130 was reported.
>> > As for now the issue was resolved and targeted for next
>> 2.1.0.0-incubating
>> > open-source release.
>> >
>> > Regards, Alex.
>> >
>> >
>> > On Mon, Jan 30, 2017 at 10:08 AM, Gagan Brahmi <ga...@gmail.com>
>> > wrote:
>> >
>> >> Hi All,
>> >>
>> >> I am not sure if anyone has faced issue for a non-privileged HAWQ user
>> >> while access hive table using HCatalog. The issue is encountered when a
>> >> non-privileged user tries to query the hive table (hive_table) under
>> any
>> >> database (default in this example) using hcatalog
>> >> (hcatalog.default.hive_table).
>> >>
>> >> However, when you describe the table (\d) there is no issue
>> encountered.
>> >>
>> >> I have tested this on HDB 2.0.0 and HDB 2.0.1.
>> >>
>> >> For the non-privileged user the following is the behavior when trying
>> to
>> >> access the data using SELECT:
>> >>
>> >> --------------------
>> >>
>> >> postgres=> SELECT * FROM hcatalog.default.hive_table;
>> >> ERROR: permission denied for relation pg_authid
>> >> LINE 1: SELECT COUNT(*) FROM hcatalog.default.hive_table;
>> >> ^
>> >> CONTEXT: SQL statement "SELECT max(oid) FROM (SELECT max(oid) AS oid
>> FROM
>> >> pg_extprotocol UNION ALL SELECT max(oid) AS oid FROM pg_partition UNION
>> >> ALL
>> >> SELECT max(oid) AS oid FROM pg_partition_rule UNION ALL SELECT
>> max(oid) AS
>> >> oid FROM pg_filespace UNION ALL SELECT max(oid) AS oid FROM
>> pg_compression
>> >> UNION ALL SELECT max(oid) AS oid FROM pg_filesystem UNION ALL SELECT
>> >> max(oid) AS oid FROM pg_foreign_data_wrapper UNION ALL SELECT max(oid)
>> AS
>> >> oid FROM pg_foreign_server UNION ALL SELECT max(oid) AS oid FROM
>> >> pg_database UNION ALL SELECT max(oid) AS oid FROM pg_type UNION ALL
>> SELECT
>> >> max(oid) AS oid FROM pg_proc UNION ALL SELECT max(oid) AS oid FROM
>> >> pg_class
>> >> UNION ALL SELECT max(oid) AS oid FROM pg_attrdef UNION ALL SELECT
>> max(oid)
>> >> AS oid FROM pg_constraint UNION ALL SELECT max(oid) AS oid FROM
>> >> pg_operator
>> >> UNION ALL SELECT max(oid) AS oid FROM pg_opclass UNION ALL SELECT
>> max(oid)
>> >> AS oid FROM pg_am UNION ALL SELECT max(oid) AS oid FROM pg_language
>> UNION
>> >> ALL SELECT max(oid) AS oid FROM pg_rewrite UNION ALL SELECT max(oid) AS
>> >> oid
>> >> FROM pg_trigger UNION ALL SELECT max(oid) AS oid FROM pg_cast UNION ALL
>> >> SELECT max(oid) AS oid FROM pg_namespace UNION ALL SELECT max(oid) AS
>> oid
>> >> FROM pg_conversion UNION ALL SELECT max(oid) AS oid FROM pg_tablespace
>> >> UNION ALL SELECT max(oid) AS oid FROM pg_resqueue UNION ALL SELECT
>> >> max(oid)
>> >> AS oid FROM pg_authid UNION ALL SELECT max(oid) AS oid FROM
>> >> pg_user_mapping) AS x"
>> >>
>> >> postgres=> SELECT * FROM hcatalog.default.hive_table;
>> >> ERROR: database does not have enough available Oids to support
>> HCatalog
>> >> queries
>> >> LINE 1: SELECT * FROM hcatalog.default.hive_table;
>> >> ^
>> >> HINT: Database VACUUM may recycle unused Oids.
>> >>
>> >>
>> >> postgres=> \d hcatalog.default.hive_table
>> >>
>> >> PXF Hive Table
>> >> "default.hive_table"
>> >> Column | Type
>> >> --------+------
>> >> id | int4
>> >> fname | text
>> >> lname | text
>> >>
>> >> --------------------
>> >>
>> >> I was able to get around this problem by granting SELECT on pg_authid
>> and
>> >> pg_user_mapping table which fixes the underlying query to grab the max
>> >> Oid.
>> >>
>> >> Is this is a known issue or is there any better workaround for this
>> >> behavior? Or anything other than granting the user SUPERUSER
>> privileges?
>> >>
>> >>
>> >> Regards,
>> >> Gagan Brahmi
>> >>
>> >
>> >
>>
>>
>>
>
>
Re: UNSUBSCRIBE
Posted by Russ Holmes <rh...@hortonworks.com>.
Unsubscribe.
----------------------------------------------------------------------
Russell Holmes
Hortonworks Public Sector
+1.703.728.0669
<https://twitter.com/zimandan>
Hadoop/DataWorks Summit<https://dataworkssummit.com/san-jose-2017/>
San Jose, CA | June 13-15
________________________________
From: Supriya Biswas <su...@nielsen.com>
Sent: Tuesday, January 31, 2017 9:03 AM
To: Contact
Subject: Re: UNSUBSCRIBE
UNSUBSCRIBE
Supriyo Biswas
Architect
The Nielsen Company
Office (516) 682-6021/NETS 249-6021
Cell (516) 353-6795
www.nielsen.com<http://www.nielsen.com/>
On Tue, Jan 31, 2017 at 8:57 AM, Brian Willingham <br...@ericsson.com>> wrote:
UNSUBSCRIBE
From: Alex (Oleksandr) Diachenko [mailto:odiachenko@pivotal.io<ma...@pivotal.io>]
Sent: Monday, January 30, 2017 1:52 PM
To: dev@hawq.incubator.apache.org<ma...@hawq.incubator.apache.org>
Cc: user@hawq.incubator.apache.org<ma...@hawq.incubator.apache.org>
Subject: Re: PXF HCatalog Issue for Non-Super User
Gagan,
Correct, that's was a mandatory condition for HCatalog intergration to work properly before the HAWQ-1130 fix.
Regards, Alex.
On Mon, Jan 30, 2017 at 11:30 AM, Gagan Brahmi <ga...@gmail.com>> wrote:
Appreciate that info Alex. Granting SELECT or SUPERUSER only viable
workaround for this problem for now. Is that a right understanding?
Regards,
Gagan Brahmi
On Mon, Jan 30, 2017 at 12:07 PM, Alex (Oleksandr) Diachenko <
odiachenko@pivotal.io<ma...@pivotal.io>> wrote:
> Hi Gagan,
>
> The issue you described was a real limitation before
> https://issues.apache.org/jira/browse/HAWQ-1130 was reported.
> As for now the issue was resolved and targeted for next 2.1.0.0-incubating
> open-source release.
>
> Regards, Alex.
>
>
> On Mon, Jan 30, 2017 at 10:08 AM, Gagan Brahmi <ga...@gmail.com>>
> wrote:
>
>> Hi All,
>>
>> I am not sure if anyone has faced issue for a non-privileged HAWQ user
>> while access hive table using HCatalog. The issue is encountered when a
>> non-privileged user tries to query the hive table (hive_table) under any
>> database (default in this example) using hcatalog
>> (hcatalog.default.hive_table).
>>
>> However, when you describe the table (\d) there is no issue encountered.
>>
>> I have tested this on HDB 2.0.0 and HDB 2.0.1.
>>
>> For the non-privileged user the following is the behavior when trying to
>> access the data using SELECT:
>>
>> --------------------
>>
>> postgres=> SELECT * FROM hcatalog.default.hive_table;
>> ERROR: permission denied for relation pg_authid
>> LINE 1: SELECT COUNT(*) FROM hcatalog.default.hive_table;
>> ^
>> CONTEXT: SQL statement "SELECT max(oid) FROM (SELECT max(oid) AS oid FROM
>> pg_extprotocol UNION ALL SELECT max(oid) AS oid FROM pg_partition UNION
>> ALL
>> SELECT max(oid) AS oid FROM pg_partition_rule UNION ALL SELECT max(oid) AS
>> oid FROM pg_filespace UNION ALL SELECT max(oid) AS oid FROM pg_compression
>> UNION ALL SELECT max(oid) AS oid FROM pg_filesystem UNION ALL SELECT
>> max(oid) AS oid FROM pg_foreign_data_wrapper UNION ALL SELECT max(oid) AS
>> oid FROM pg_foreign_server UNION ALL SELECT max(oid) AS oid FROM
>> pg_database UNION ALL SELECT max(oid) AS oid FROM pg_type UNION ALL SELECT
>> max(oid) AS oid FROM pg_proc UNION ALL SELECT max(oid) AS oid FROM
>> pg_class
>> UNION ALL SELECT max(oid) AS oid FROM pg_attrdef UNION ALL SELECT max(oid)
>> AS oid FROM pg_constraint UNION ALL SELECT max(oid) AS oid FROM
>> pg_operator
>> UNION ALL SELECT max(oid) AS oid FROM pg_opclass UNION ALL SELECT max(oid)
>> AS oid FROM pg_am UNION ALL SELECT max(oid) AS oid FROM pg_language UNION
>> ALL SELECT max(oid) AS oid FROM pg_rewrite UNION ALL SELECT max(oid) AS
>> oid
>> FROM pg_trigger UNION ALL SELECT max(oid) AS oid FROM pg_cast UNION ALL
>> SELECT max(oid) AS oid FROM pg_namespace UNION ALL SELECT max(oid) AS oid
>> FROM pg_conversion UNION ALL SELECT max(oid) AS oid FROM pg_tablespace
>> UNION ALL SELECT max(oid) AS oid FROM pg_resqueue UNION ALL SELECT
>> max(oid)
>> AS oid FROM pg_authid UNION ALL SELECT max(oid) AS oid FROM
>> pg_user_mapping) AS x"
>>
>> postgres=> SELECT * FROM hcatalog.default.hive_table;
>> ERROR: database does not have enough available Oids to support HCatalog
>> queries
>> LINE 1: SELECT * FROM hcatalog.default.hive_table;
>> ^
>> HINT: Database VACUUM may recycle unused Oids.
>>
>>
>> postgres=> \d hcatalog.default.hive_table
>>
>> PXF Hive Table
>> "default.hive_table"
>> Column | Type
>> --------+------
>> id | int4
>> fname | text
>> lname | text
>>
>> --------------------
>>
>> I was able to get around this problem by granting SELECT on pg_authid and
>> pg_user_mapping table which fixes the underlying query to grab the max
>> Oid.
>>
>> Is this is a known issue or is there any better workaround for this
>> behavior? Or anything other than granting the user SUPERUSER privileges?
>>
>>
>> Regards,
>> Gagan Brahmi
>>
>
>
Re: UNSUBSCRIBE
Posted by Supriya Biswas <su...@nielsen.com>.
UNSUBSCRIBE
*Supriyo Biswas *Architect
The Nielsen Company
Office (516) 682-6021/NETS 249-6021
Cell (516) 353-6795
www.nielsen.com
On Tue, Jan 31, 2017 at 8:57 AM, Brian Willingham <
brian.willingham@ericsson.com> wrote:
> UNSUBSCRIBE
>
>
>
>
>
> *From:* Alex (Oleksandr) Diachenko [mailto:odiachenko@pivotal.io]
> *Sent:* Monday, January 30, 2017 1:52 PM
> *To:* dev@hawq.incubator.apache.org
> *Cc:* user@hawq.incubator.apache.org
> *Subject:* Re: PXF HCatalog Issue for Non-Super User
>
>
>
> Gagan,
>
>
>
> Correct, that's was a mandatory condition for HCatalog intergration to
> work properly before the HAWQ-1130 fix.
>
>
>
> Regards, Alex.
>
>
>
> On Mon, Jan 30, 2017 at 11:30 AM, Gagan Brahmi <ga...@gmail.com>
> wrote:
>
> Appreciate that info Alex. Granting SELECT or SUPERUSER only viable
> workaround for this problem for now. Is that a right understanding?
>
>
> Regards,
> Gagan Brahmi
>
>
> On Mon, Jan 30, 2017 at 12:07 PM, Alex (Oleksandr) Diachenko <
> odiachenko@pivotal.io> wrote:
>
> > Hi Gagan,
> >
> > The issue you described was a real limitation before
> > https://issues.apache.org/jira/browse/HAWQ-1130 was reported.
> > As for now the issue was resolved and targeted for next
> 2.1.0.0-incubating
> > open-source release.
> >
> > Regards, Alex.
> >
> >
> > On Mon, Jan 30, 2017 at 10:08 AM, Gagan Brahmi <ga...@gmail.com>
> > wrote:
> >
> >> Hi All,
> >>
> >> I am not sure if anyone has faced issue for a non-privileged HAWQ user
> >> while access hive table using HCatalog. The issue is encountered when a
> >> non-privileged user tries to query the hive table (hive_table) under any
> >> database (default in this example) using hcatalog
> >> (hcatalog.default.hive_table).
> >>
> >> However, when you describe the table (\d) there is no issue encountered.
> >>
> >> I have tested this on HDB 2.0.0 and HDB 2.0.1.
> >>
> >> For the non-privileged user the following is the behavior when trying to
> >> access the data using SELECT:
> >>
> >> --------------------
> >>
> >> postgres=> SELECT * FROM hcatalog.default.hive_table;
> >> ERROR: permission denied for relation pg_authid
> >> LINE 1: SELECT COUNT(*) FROM hcatalog.default.hive_table;
> >> ^
> >> CONTEXT: SQL statement "SELECT max(oid) FROM (SELECT max(oid) AS oid
> FROM
> >> pg_extprotocol UNION ALL SELECT max(oid) AS oid FROM pg_partition UNION
> >> ALL
> >> SELECT max(oid) AS oid FROM pg_partition_rule UNION ALL SELECT max(oid)
> AS
> >> oid FROM pg_filespace UNION ALL SELECT max(oid) AS oid FROM
> pg_compression
> >> UNION ALL SELECT max(oid) AS oid FROM pg_filesystem UNION ALL SELECT
> >> max(oid) AS oid FROM pg_foreign_data_wrapper UNION ALL SELECT max(oid)
> AS
> >> oid FROM pg_foreign_server UNION ALL SELECT max(oid) AS oid FROM
> >> pg_database UNION ALL SELECT max(oid) AS oid FROM pg_type UNION ALL
> SELECT
> >> max(oid) AS oid FROM pg_proc UNION ALL SELECT max(oid) AS oid FROM
> >> pg_class
> >> UNION ALL SELECT max(oid) AS oid FROM pg_attrdef UNION ALL SELECT
> max(oid)
> >> AS oid FROM pg_constraint UNION ALL SELECT max(oid) AS oid FROM
> >> pg_operator
> >> UNION ALL SELECT max(oid) AS oid FROM pg_opclass UNION ALL SELECT
> max(oid)
> >> AS oid FROM pg_am UNION ALL SELECT max(oid) AS oid FROM pg_language
> UNION
> >> ALL SELECT max(oid) AS oid FROM pg_rewrite UNION ALL SELECT max(oid) AS
> >> oid
> >> FROM pg_trigger UNION ALL SELECT max(oid) AS oid FROM pg_cast UNION ALL
> >> SELECT max(oid) AS oid FROM pg_namespace UNION ALL SELECT max(oid) AS
> oid
> >> FROM pg_conversion UNION ALL SELECT max(oid) AS oid FROM pg_tablespace
> >> UNION ALL SELECT max(oid) AS oid FROM pg_resqueue UNION ALL SELECT
> >> max(oid)
> >> AS oid FROM pg_authid UNION ALL SELECT max(oid) AS oid FROM
> >> pg_user_mapping) AS x"
> >>
> >> postgres=> SELECT * FROM hcatalog.default.hive_table;
> >> ERROR: database does not have enough available Oids to support HCatalog
> >> queries
> >> LINE 1: SELECT * FROM hcatalog.default.hive_table;
> >> ^
> >> HINT: Database VACUUM may recycle unused Oids.
> >>
> >>
> >> postgres=> \d hcatalog.default.hive_table
> >>
> >> PXF Hive Table
> >> "default.hive_table"
> >> Column | Type
> >> --------+------
> >> id | int4
> >> fname | text
> >> lname | text
> >>
> >> --------------------
> >>
> >> I was able to get around this problem by granting SELECT on pg_authid
> and
> >> pg_user_mapping table which fixes the underlying query to grab the max
> >> Oid.
> >>
> >> Is this is a known issue or is there any better workaround for this
> >> behavior? Or anything other than granting the user SUPERUSER privileges?
> >>
> >>
> >> Regards,
> >> Gagan Brahmi
> >>
> >
> >
>
>
>