Lists: | Postg사설 토토SQL |
---|
From: | Glen Barber <glen(dot)j(dot)barber(at)gmail(dot)com> |
---|---|
To: | pgsql-testers(at)postgresql(dot)org |
Subject: | pg_upgrade and reg* system oid user types |
Date: | 2010-08-03 17:07:03 |
Message-ID: | 4C584CB7.6030500@gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Lists: | Postg롤 토토SQL : |
Hi,
I think it might be worth mentioning in the release notes (unless I've
missed it, of course) that, with the addition of
check_for_reg_data_type_usage(), pg_upgrade will error if it finds, for
example, tsearch data from pre-8.3 in user tables.
Otherwise, pg_upgrade has worked wonderfully for me - thanks again for
bringing it into contrib.
Best regards,
--
Glen Barber
From: | Josh Berkus <josh(at)agliodbs(dot)com> |
---|---|
To: | Glen Barber <glen(dot)j(dot)barber(at)gmail(dot)com> |
Cc: | pgsql-testers(at)postgresql(dot)org |
Subject: | Re: pg_upgrade and reg* system oid user types |
Date: | 2010-08-03 18:18:01 |
Message-ID: | 4C585D59.20403@agliodbs.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Lists: | pgsql-testers |
On 8/3/10 10:07 AM, Glen Barber wrote:
> Hi,
>
> I think it might be worth mentioning in the release notes (unless I've
> missed it, of course) that, with the addition of
> check_for_reg_data_type_usage(), pg_upgrade will error if it finds, for
> example, tsearch data from pre-8.3 in user tables.
>
> Otherwise, pg_upgrade has worked wonderfully for me - thanks again for
> bringing it into contrib.
Good to hear. Problem report forwarded to Bruce.
--
-- Josh Berkus
PostgreSQL Experts Inc.
http://www.pgexperts.com
From: | Bruce Momjian <bruce(at)momjian(dot)us> |
---|---|
To: | Josh Berkus <josh(at)agliodbs(dot)com> |
Cc: | Glen Barber <glen(dot)j(dot)barber(at)gmail(dot)com>, pgsql-testers(at)postgresql(dot)org |
Subject: | Re: pg_upgrade and reg* system oid user types |
Date: | 2010-08-03 18:19:41 |
Message-ID: | 201008031819.o73IJfU23806@momjian.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Lists: | Postg토토 핫SQL : |
Josh Berkus wrote:
> On 8/3/10 10:07 AM, Glen Barber wrote:
> > Hi,
> >
> > I think it might be worth mentioning in the release notes (unless I've
> > missed it, of course) that, with the addition of
> > check_for_reg_data_type_usage(), pg_upgrade will error if it finds, for
> > example, tsearch data from pre-8.3 in user tables.
> >
> > Otherwise, pg_upgrade has worked wonderfully for me - thanks again for
> > bringing it into contrib.
>
> Good to hear. Problem report forwarded to Bruce.
I was wondering if users were storing text search reg* data types in
their user tables. Is this common? Why only pre-8.3?
--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com
+ It's impossible for everything to be true. +
From: | Bruce Momjian <bruce(at)momjian(dot)us> |
---|---|
To: | Glen Barber <glen(dot)j(dot)barber(at)gmail(dot)com> |
Cc: | pgsql-testers(at)postgresql(dot)org |
Subject: | Re: pg_upgrade and reg* system oid user types |
Date: | 2010-08-03 18:27:15 |
Message-ID: | 201008031827.o73IRFw24984@momjian.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Lists: | Postg사설 토토SQL |
Glen Barber wrote:
> Hi,
>
> I think it might be worth mentioning in the release notes (unless I've
> missed it, of course) that, with the addition of
> check_for_reg_data_type_usage(), pg_upgrade will error if it finds, for
> example, tsearch data from pre-8.3 in user tables.
This is already mentioned in the pg_upgrade manual page. Since this is
a new utility, and hence not changed from a preivous release, it would
not be appropriate to mention this restriction in the release notes,
unless it is a restriction we find to be common.
--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com
+ It's impossible for everything to be true. +
From: | Glen Barber <glen(dot)j(dot)barber(at)gmail(dot)com> |
---|---|
To: | Bruce Momjian <bruce(at)momjian(dot)us> |
Cc: | Josh Berkus <josh(at)agliodbs(dot)com>, pgsql-testers(at)postgresql(dot)org |
Subject: | Re: pg_upgrade and reg* system oid user types |
Date: | 2010-08-03 18:29:40 |
Message-ID: | 4C586014.7030900@gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Lists: | pgsql-testers |
On 8/3/10 2:19 PM, Bruce Momjian wrote:
> Josh Berkus wrote:
>> On 8/3/10 10:07 AM, Glen Barber wrote:
>>> Hi,
>>>
>>> I think it might be worth mentioning in the release notes (unless I've
>>> missed it, of course) that, with the addition of
>>> check_for_reg_data_type_usage(), pg_upgrade will error if it finds, for
>>> example, tsearch data from pre-8.3 in user tables.
>>>
>>> Otherwise, pg_upgrade has worked wonderfully for me - thanks again for
>>> bringing it into contrib.
>>
>> Good to hear. Problem report forwarded to Bruce.
>
> I was wondering if users were storing text search reg* data types in
> their user tables. Is this common? Why only pre-8.3?
>
The database I am using in testing 9.0 is a development database which
mimics our production environment as closely as it can. This particular
data has been through several PostgreSQL upgrades, beginning with 7.4,
if I remember correctly. The reg* data types in user tables is from
before tsearch functionality was added to core.
Regards,
--
Glen Barber
From: | Bruce Momjian <bruce(at)momjian(dot)us> |
---|---|
To: | Glen Barber <glen(dot)j(dot)barber(at)gmail(dot)com> |
Cc: | Josh Berkus <josh(at)agliodbs(dot)com>, pgsql-testers(at)postgresql(dot)org |
Subject: | Re: pg_upgrade and reg* system oid user types |
Date: | 2010-08-03 18:36:07 |
Message-ID: | 201008031836.o73Ia7026565@momjian.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Lists: | pgsql-testers |
Glen Barber wrote:
> On 8/3/10 2:19 PM, Bruce Momjian wrote:
> > Josh Berkus wrote:
> >> On 8/3/10 10:07 AM, Glen Barber wrote:
> >>> Hi,
> >>>
> >>> I think it might be worth mentioning in the release notes (unless I've
> >>> missed it, of course) that, with the addition of
> >>> check_for_reg_data_type_usage(), pg_upgrade will error if it finds, for
> >>> example, tsearch data from pre-8.3 in user tables.
> >>>
> >>> Otherwise, pg_upgrade has worked wonderfully for me - thanks again for
> >>> bringing it into contrib.
> >>
> >> Good to hear. Problem report forwarded to Bruce.
> >
> > I was wondering if users were storing text search reg* data types in
> > their user tables. Is this common? Why only pre-8.3?
> >
>
> The database I am using in testing 9.0 is a development database which
> mimics our production environment as closely as it can. This particular
> data has been through several PostgreSQL upgrades, beginning with 7.4,
> if I remember correctly. The reg* data types in user tables is from
> before tsearch functionality was added to core.
OK, good to know. I am hoping that isn't something that many people
have in their databases. We can fix the reg* migration, but it is too
late for 9.0 and fixing it will require additional code in the backend
that we would prefer to avoid.
--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com
+ It's impossible for everything to be true. +
From: | Glen Barber <glen(dot)j(dot)barber(at)gmail(dot)com> |
---|---|
To: | Bruce Momjian <bruce(at)momjian(dot)us> |
Cc: | Josh Berkus <josh(at)agliodbs(dot)com>, pgsql-testers(at)postgresql(dot)org |
Subject: | Re: pg_upgrade and reg* system oid user types |
Date: | 2010-08-03 18:46:46 |
Message-ID: | 4C586416.1050608@gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Lists: | pgsql-testers |
On 8/3/10 2:36 PM, Bruce Momjian wrote:
>>> I was wondering if users were storing text search reg* data types in
>>> their user tables. Is this common? Why only pre-8.3?
>>>
>>
>> The database I am using in testing 9.0 is a development database which
>> mimics our production environment as closely as it can. This particular
>> data has been through several PostgreSQL upgrades, beginning with 7.4,
>> if I remember correctly. The reg* data types in user tables is from
>> before tsearch functionality was added to core.
>
> OK, good to know. I am hoping that isn't something that many people
> have in their databases. We can fix the reg* migration, but it is too
> late for 9.0 and fixing it will require additional code in the backend
> that we would prefer to avoid.
>
I wasn't sure if it would be common. I was able to move forward with a
dump/restore for now; I'll remove the user tables in question and run it
through pg_upgrade again later this week.
Thanks for the feedback.
Regards,
--
Glen Barber
From: | Bruce Momjian <bruce(at)momjian(dot)us> |
---|---|
To: | Glen Barber <glen(dot)j(dot)barber(at)gmail(dot)com> |
Cc: | Josh Berkus <josh(at)agliodbs(dot)com>, pgsql-testers(at)postgresql(dot)org |
Subject: | Re: pg_upgrade and reg* system oid user types |
Date: | 2010-08-03 19:05:06 |
Message-ID: | 201008031905.o73J56F02178@momjian.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Lists: | pgsql-testers |
Glen Barber wrote:
> On 8/3/10 2:36 PM, Bruce Momjian wrote:
> >>> I was wondering if users were storing text search reg* data types in
> >>> their user tables. Is this common? Why only pre-8.3?
> >>>
> >>
> >> The database I am using in testing 9.0 is a development database which
> >> mimics our production environment as closely as it can. This particular
> >> data has been through several PostgreSQL upgrades, beginning with 7.4,
> >> if I remember correctly. The reg* data types in user tables is from
> >> before tsearch functionality was added to core.
> >
> > OK, good to know. I am hoping that isn't something that many people
> > have in their databases. We can fix the reg* migration, but it is too
> > late for 9.0 and fixing it will require additional code in the backend
> > that we would prefer to avoid.
> >
>
> I wasn't sure if it would be common. I was able to move forward with a
> dump/restore for now; I'll remove the user tables in question and run it
> through pg_upgrade again later this week.
Yes, pg_upgrade should have recommended manual upgrade of tables using
reg* so you could still use pg_upgrade.
--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com
+ It's impossible for everything to be true. +