It's strange. I can reproduce first query issue from Dashboard Data Viewer.

On second query, I have about 10k record in the table and get about
half of them using cursor operation like following.

time = memcache.get('cursor')
users = Users.all().order('-updated_at').filter('updated_at <', time).fetch(100)
memcache.set('cursor', users[-1].updated_at)

And yes, all timeout happens in datastore, not deadline exceeded.

On Thu, Oct 1, 2009 at 10:37 AM, Jeff S (Google) <j...@google.com> wrote:
> Ah I see, thank you for the details. So far I haven't been able to reproduce
> the timeout on the first query that you listed. How are you performing the
> query and how large are these entities? It may be that the cause of the
> timeout is the way in which the query is being performed. Also, I wanted to
> confirm that this a datastore timeout (as opposed to deadline exceeded).
>
> Thank you,
>
> Jeff
>
> On Wed, Sep 30, 2009 at 9:01 PM, Kazuho Okui <kaz...@gmail.com> wrote:
>>
>> SELECT * FROM User order by updated_at
>>
>> This query always returns timeout. It doesn't matter query from python
>> or from dashboard.
>>
>> SELECT * FROM User where order by updated_at desc
>>
>> This one works better, however, it still gets a lot of errors. (error
>> rate is  5-10%)  I query 100 records per one http request. Is this
>> usual error rate?
>>
>> On Wed, Sep 30, 2009 at 2:23 PM, Jeff S (Google) <j...@google.com> wrote:
>> > Hi Kazuho Okui,
>> >
>> > Which queries are giving you errors? It looks like you have an index on
>> > updated_at which is ascending and if there are no other filters in your
>> > query with updated_at descending then the single property indexes (there
>> > are
>> > two for each indexed property, one ascending, one descending) should be
>> > able
>> > to support this query. What is the specific error that you are seeing?
>> >
>> > Thank you,
>> >
>> > Jeff
>> >
>> > On Wed, Sep 30, 2009 at 10:02 AM, Kazuho Okui <kaz...@gmail.com> wrote:
>> >>
>> >> How long will this take? Fixing just single index issue takes more than
>> >> 2
>> >> weeks?
>> >>
>> >> On Tue, Sep 22, 2009 at 11:28 PM, Kazuho Okui <kaz...@gmail.com> wrote:
>> >> > It seems a lot of errors occur in 'order by updated_at desc' now.
>> >> > 'order by updated_at' doesn't work for a week.
>> >> >
>> >> > Any updates?
>> >> >
>> >> > On Sat, Sep 19, 2009 at 12:16 PM, Kazuho Okui <kaz...@gmail.com>
>> >> > wrote:
>> >> >> Thanks! I removed stuck indexes. However, I still get errors on
>> >> >> 'order
>> >> >> by updated_at' even 'order by updated_at desc' sometimes.
>> >> >>
>> >> >> - Kazuho
>> >> >>
>> >> >> On Fri, Sep 18, 2009 at 11:05 AM, Jeff S (Google) <j...@google.com>
>> >> >> wrote:
>> >> >>> Hi Kazuho,
>> >> >>>
>> >> >>> I've set the state of the stuck indexes to error for
>> >> >>> twitrackbacktest
>> >> >>> so
>> >> >>> that they can be removed. I'm still looking into possible issues in
>> >> >>> echofonsync.
>> >> >>>
>> >> >>> Thank you,
>> >> >>>
>> >> >>> Jeff
>> >> >>>
>> >> >>> On Fri, Sep 18, 2009 at 10:23 AM, Kazuho Okui <kaz...@gmail.com>
>> >> >>> wrote:
>> >> >>>>
>> >> >>>> Hi
>> >> >>>>
>> >> >>>> I have 2 instance.
>> >> >>>>
>> >> >>>> app_id: twitrackbacktest
>> >> >>>>
>> >> >>>> Indexing stuck. Building / deleting index stuck. it takes almost
>> >> >>>> 2weeks.
>> >> >>>>
>> >> >>>>
>> >> >>>> app_id: echofonsync
>> >> >>>>
>> >> >>>> index of 'updated_at' is somehow broken. Following query always
>> >> >>>> returns datastore_errors.timeout
>> >> >>>>
>> >> >>>> SELECT * FROM Users ORDER BY updated_at
>> >> >>>>
>> >> >>>> related query also causes error even I re-create index.
>> >> >>>>
>> >> >>>> SELECT * FROM User ORDER BY updated_at,device_token
>> >> >>>>
>> >> >>>> It's very odd, but if I add 'desc' to updated_at, it works.
>> >> >>>>
>> >> >>>> SELECT * FROM User ORDER BY updated_at desc
>> >> >>>>
>> >> >>>> Other indexes works fine.
>> >> >>>>
>> >> >>>> SELECT * FROM User ORDER BY user_id
>> >> >>>> SELECT * FROM User ORDER BY username
>> >> >>>> ...
>> >> >>>>
>> >> >>>>
>> >> >>>> - Kazuho
>> >> >>>>
>> >> >>>>
>> >> >>>
>> >> >>>
>> >> >>> >>>
>> >> >>>
>> >> >>
>> >> >
>> >>
>> >>
>> >
>> >
>> > >
>> >
>>
>>
>
>
> >
>

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"Google App Engine" group.
To post to this group, send email to google-appengine@googlegroups.com
To unsubscribe from this group, send email to 
google-appengine+unsubscr...@googlegroups.com
For more options, visit this group at 
http://groups.google.com/group/google-appengine?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to