[google-appengine] Re: App Engine HTTP issues

[Krisztián Guba]

Hi,
Yes you are right, that was the case, but after a incredibly long 
deployment I could solve the situation.
It looks working fine now.
Thanks,
Krisztián Guba

On Wednesday, 7 October 2020 at 22:29:31 UTC+2 hi...@google.com wrote:

> Hello,
>
> If I understand correctly, the original version of your app engine 
> application which previously threw "java.net.UnknownHostException, 
> ServiceConfigException: java.net.UnknownHostException: 
> servicemanagement.googleapis.com" still fails to deploy. Also, if I 
> understood correctly after migrating traffic to the "test" version of the 
> service, traffic is still occasionally routed to the master version.
>
> If this is indeed the case, it is likely that the issue with App Engine 
> may have some residual effects on the affected service, and thus I would 
> recommend creating a new Public Tracker via the Issue Tracker 
>  so that Support 
> Representatives can investigate the cause accordingly. 
>
> On Wednesday, October 7, 2020 at 9:13:49 AM UTC-4 
> kris...@escapemasters.co.nz wrote:
>
>> So, I cannot deploy to my master version and traffic migration still 
>> fails.
>>
>> On Tuesday, 6 October 2020 at 20:33:53 UTC+2 Katayoon (Cloud Platform 
>> Support) wrote:
>>
>>> Hi,
>>>
>>> The issue with Google App Engine 
>>>  is now 
>>> resolved. 
>>>
>>> On Tuesday, October 6, 2020 at 11:39:27 AM UTC-4 goo...@ryanb.org wrote:
>>>
 Looks like they're in the middle of a partial outage right now. 
 https://status.cloud.google.com/

 On Tuesday, October 6, 2020 at 8:08:44 AM UTC-7 
 kris...@escapemasters.co.nz wrote:

> Hi All,
> In the past hour we faced on standard java App Engine with numerous 
> unexpected error as java.net.UnknownHostException, 
> ServiceConfigException: 
> java.net.UnknownHostException: servicemanagement.googleapis.com. On 
> our side nothing has been changed. 
> Moreover, when I tried to migrate the traffic to an older version the 
> traffic migration failed in half way.
>
> Who knows anything about this kind of outage?
>
> Thanks in advance,
> Krisztián Guba
>


-- 
You received this message because you are subscribed to the Google Groups 
"Google App Engine" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to google-appengine+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/google-appengine/aa0ff940-5858-4de8-8e19-76b598b2cdbfn%40googlegroups.com.

Re: [google-appengine] Uploads to Legacy Blobstore are Failing with 500, nothing in the logs

[Joshua Smith]

Indeed, there is no such account with any permission on that bucket.

So there you have it. That’s probably what happened. (Doesn’t matter anymore 
for me, since I’m now using cloudstorage for uploads, not blobstore.)

Google really botched this transition from appcfg to gcloud.

-Joshua

> On Oct 9, 2020, at 4:42 AM, 'Jose V' via Google App Engine 
>  wrote:
> 
> Hello,
> 
> I have seen previous occurrences in which deploying with "gcloud app deploy" 
> led to 50X errors when uploading to blobstore. It was indeed due to the App 
> Engine Service account lacking permissions on the newly created Cloud Storage 
> bucket.
> Note that by default, the Service account used by App Engine has a name like 
> "[PROJECT-ID]@appspot.gserviceaccount.com" and the Cloud Storage Bucket will 
> have a name like "[PROJECT-ID].appspot.com "
> 
> Could you confirm that the App Engine service account has write permissions 
> for the bucket? See Cloud Storage Roles [1] for reference.
> 
> Regards
> [1]: https://cloud.google.com/storage/docs/access-control/iam-roles 
> 
> 
> On Thursday, October 8, 2020 at 7:23:05 PM UTC+2 Will Reiher wrote:
> I don't remember what the error code was anymore. I just knew that something 
> was f'd. It was really a hunch that that new bucket was involved and that if 
> my app didn't have permission it might be the answer. I was prepared for a 
> very long day but after the permission change it all just started working 
> again - with the blobs showing up in that new bucket.
> 
> Most of my stuff is using Cloud Storage except for user uploads. I'm working 
> to replace that with a UUID based storage name so that I can maintain the 
> ability to allow for same-named files to be uploaded.
> 
> Of course today Cloud Task UI is not available and it wants me to enable the 
> API (I've been using it for months now) Nervous that my queues will be erased 
> if I enable it.
> 
> On Thursday, October 8, 2020 at 10:03:22 AM UTC-7 Joshua Smith wrote:
> Do you recall if you got 500 errors? I’d expect that to generate 403’s.
> 
> Regardless, that does sound like the kind of thing that happened.
> 
> Switching to use cloudstorage instead of blobstore wasn’t too hard, so that’s 
> the workaround I’d recommend for anyone else who hits this issue.
> 
> I’m mostly just annoyed that google has switched to a support model where 
> they won’t provide timely support unless you pay them $100/month. That’s 
> greedy and stupid.
> 
> -Joshua
> 
> 
>> On Oct 8, 2020, at 12:58 PM, Will Reiher > 
>> wrote:
>> 
> 
>> I had an issue a few weeks back. It was due to my transition to gcloud 
>> deploy. It created a new app-id.appspot.com  
>> bucket but my app engine service account did not have permissions. Once I 
>> gave it permission it started working again with the blobstore items showing 
>> up in the new "default" bucket.
>> 
>> On Thursday, October 8, 2020 at 3:13:28 AM UTC-7 barrado wrote:
>> 
>> Hi,
>> 
>> I have deployed this Blobstore API sample 
>> 
>>  service and it works well and allows me to successfully upload images. You 
>> could try to add further logging statements in your app to know what is 
>> causing your issue. 
>> 
>> On the other hand, I recommend you to fully migrate you app to Python3 since 
>> Python 2 is no longer supported by the community and use Cloud Storage 
>> instead of Blobstore.
>> 
>> And also note that appcfg tooling was shuted down 
>>  on August 30th, 2020.
>> 
>> 
>> On Wednesday, October 7, 2020 at 4:17:44 PM UTC+2 Joshua Smith wrote:
>> Unfortunately, I used gcloud app deploy because appcfg.py stopped working a 
>> couple weeks ago.
>> 
>> I did use curl and it revealed nothing. Just a terse 500 with no content 
>> from the server.
>> 
>> 
>>> On Oct 6, 2020, at 4:14 PM, Charounson Saintilus > 
>>> wrote:
>>> 
>> 
>>> Hey Joshua,
>>> 
>>> I don’t think I’m gonna be of much help here, but wanted to mention: if you 
>>> think that running “gcloud app deploy” introduced the issue, can you try 
>>> redeploying the application using appcfg.py? 
>>> 
>>> You could also use curl commands to troubleshoot the upload to blobstore.. 
>>> 
>>> Thanks,
>>> 
>>> Char 
>>> 
>>> On Tue, Oct 6, 2020 at 3:40 PM Joshua Smith > wrote:
>>> The problem persists.
>>> 
>>> It occurs to me that last week was the first time I had to make a small 
>>> update to the app in a while, and I did that using “gcloud app deploy”  
>>> instead of appcfg.py for the first time on that project.
>>> 
>>> Is it possible that pushing the app using gcloud introduced a problem 
>>> uploading to blobstore? I wouldn’t think they’d be related, but the timing 
>>> is suspicious.
>>> 
>>> Downloading from blobstore still works. It’s only uploads that are failing.

[google-appengine] Datastore Enum Types No Longer Serialized

[Larry Schiefer]

We are using GAE python standard environment (Python 2.7) and suddenly our 
app started experiencing issues when serializing a datastore enum property. 
Our code has been deployed and working for over 6 months and today it is 
failing. For example:

class SomeType(messages.Enum):
FIRST_TYPE = 1
SECOND_TYPE = 2
THIRD_TYPE = 3

class SomeEntity(polymodel.PolyModel):
# Other fields not shown
something = msgprop.EnumProperty(SomeType, repeated=true)

def get_msg_data(self):
msg_data = {
'something': json.dumps(self.something)
}
return msg_data

The call to json.dumps above started failing in our instance today with an 
error like this:

TypeError: SomeType(FIRST_TYPE, 1) is not JSON serializable

Is anyone else seeing this? Note that using the same enum type within a 
message where we are not calling json.dumps directly works fine. 

-- 
You received this message because you are subscribed to the Google Groups 
"Google App Engine" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to google-appengine+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/google-appengine/07b1be86-920a-4ff5-8a78-2a4a840273efn%40googlegroups.com.

[google-appengine] Datastore emulator message type mismatch

[Jagath Weerasinghe]

Hello, 

I tried to start the datastore emulator in my local dev environment (Java 
8) and got the following error. Looks to me it happens because of a 
mismatch of protobuf versions. It would be great if you know of any 
workaround for this. Thanks. 

Jagath

Cloud SDK Version: 313. 
OS: macOS

$ gcloud beta emulators datastore start

Executing: 
/Users/w/Work/google-cloud/google-cloud-sdk/platform/cloud-datastore-emulator/cloud_datastore_emulator
 
create --project_id= /Users/w/.config/gcloud/emulators/datastore

[datastore] Oct 09, 2020 12:28:45 PM 
com.google.cloud.datastore.emulator.CloudDatastore$CreateAction$1 apply

[datastore] INFO: Provided project_id to Cloud Datastore emulator create 
command, which is no longer necessary.

[datastore] Created new Cloud Datastore project in 
'/Users/w/.config/gcloud/emulators/datastore'.

Executing: 
/Users/w/Work/google-cloud/google-cloud-sdk/platform/cloud-datastore-emulator/cloud_datastore_emulator
 
start --host=localhost --port=8081 --store_on_disk=True --consistency=0.9 
--allow_remote_shutdown /Users/w/.config/gcloud/emulators/datastore

[datastore] Oct 09, 2020 12:28:45 PM 
com.google.cloud.datastore.emulator.CloudDatastore$FakeDatastoreAction$9 
apply

[datastore] INFO: Provided --allow_remote_shutdown to start command which 
is no longer necessary.

[datastore] Oct 09, 2020 12:28:45 PM 
com.google.cloud.datastore.emulator.impl.LocalDatastoreFileStub 

[datastore] INFO: Local Datastore initialized:

[datastore] Type: High Replication

[datastore] Storage: 
/Users/w/.config/gcloud/emulators/datastore/WEB-INF/appengine-generated/local_db.bin

[datastore] *Exiting due to exception: java.lang.IllegalArgumentException: 
Extension is for type "proto2.MethodOptions" which does not match message 
type "google.protobuf.MethodOptions".*

-- 
You received this message because you are subscribed to the Google Groups 
"Google App Engine" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to google-appengine+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/google-appengine/0b7f6549-761f-409a-8960-9f94d9715f5en%40googlegroups.com.

Re: [google-appengine] Uploads to Legacy Blobstore are Failing with 500, nothing in the logs

['Jose V' via Google App Engine]

Hello,

I have seen previous occurrences in which deploying with "gcloud app 
deploy" led to 50X errors when uploading to blobstore. It was indeed due to 
the App Engine Service account lacking permissions on the newly created 
Cloud Storage bucket.
Note that by default, the Service account used by App Engine has a name 
like "[PROJECT-ID]@appspot.gserviceaccount.com" and the Cloud Storage 
Bucket will have a name like "[PROJECT-ID].appspot.com"

Could you confirm that the App Engine service account has write permissions 
for the bucket? See Cloud Storage Roles [1] for reference.

Regards
[1]: https://cloud.google.com/storage/docs/access-control/iam-roles

On Thursday, October 8, 2020 at 7:23:05 PM UTC+2 Will Reiher wrote:

> I don't remember what the error code was anymore. I just knew that 
> something was f'd. It was really a hunch that that new bucket was involved 
> and that if my app didn't have permission it might be the answer. I was 
> prepared for a very long day but after the permission change it all just 
> started working again - with the blobs showing up in that new bucket.
>
> Most of my stuff is using Cloud Storage except for user uploads. I'm 
> working to replace that with a UUID based storage name so that I can 
> maintain the ability to allow for same-named files to be uploaded.
>
> Of course today Cloud Task UI is not available and it wants me to enable 
> the API (I've been using it for months now) Nervous that my queues will be 
> erased if I enable it.
>
> On Thursday, October 8, 2020 at 10:03:22 AM UTC-7 Joshua Smith wrote:
>
>> Do you recall if you got 500 errors? I’d expect that to generate 403’s.
>>
>> Regardless, that does sound like the kind of thing that happened.
>>
>> Switching to use cloudstorage instead of blobstore wasn’t too hard, so 
>> that’s the workaround I’d recommend for anyone else who hits this issue.
>>
>> I’m mostly just annoyed that google has switched to a support model where 
>> they won’t provide timely support unless you pay them $100/month. That’s 
>> greedy and stupid.
>>
>> -Joshua
>>
>> On Oct 8, 2020, at 12:58 PM, Will Reiher  
>> wrote:
>>
>> I had an issue a few weeks back. It was due to my transition to gcloud 
>> deploy. It created a new app-id.appspot.com bucket but my app engine 
>> service account did not have permissions. Once I gave it permission it 
>> started working again with the blobstore items showing up in the new 
>> "default" bucket.
>>
>> On Thursday, October 8, 2020 at 3:13:28 AM UTC-7 barrado wrote:
>>
>>>
>>> Hi,
>>>
>>> I have deployed this Blobstore API sample 
>>> 
>>>  service 
>>> and it works well and allows me to successfully upload images. You could 
>>> try to add further logging statements in your app to know what is causing 
>>> your issue. 
>>>
>>> On the other hand, I recommend you to fully migrate you app to Python3 
>>> since Python 2 is no longer supported by the community and use Cloud 
>>> Storage instead of Blobstore.
>>>
>>> And also note that appcfg tooling was shuted down 
>>>  on August 30th, 
>>> 2020.
>>>
>>>
>>> On Wednesday, October 7, 2020 at 4:17:44 PM UTC+2 Joshua Smith wrote:
>>>
 Unfortunately, I used gcloud app deploy because appcfg.py stopped 
 working a couple weeks ago.

 I did use curl and it revealed nothing. Just a terse 500 with no 
 content from the server.

 On Oct 6, 2020, at 4:14 PM, Charounson Saintilus  
 wrote:

 Hey Joshua,

 I don’t think I’m gonna be of much help here, but wanted to mention: if 
 you think that running “gcloud app deploy” introduced the issue, can 
 you try redeploying the application using appcfg.py? 

 You could also use curl commands to troubleshoot the upload to 
 blobstore.. 

 Thanks,

 Char 

 On Tue, Oct 6, 2020 at 3:40 PM Joshua Smith  
 wrote:

> The problem persists.
>
> It occurs to me that last week was the first time I had to make a 
> small update to the app in a while, and I did that using “gcloud app 
> deploy”  instead of appcfg.py for the first time on that project.
>
> Is it possible that pushing the app using gcloud introduced a problem 
> uploading to blobstore? I wouldn’t think they’d be related, but the 
> timing 
> is suspicious.
>
> Downloading from blobstore still works. It’s only uploads that are 
> failing.
>
> My users are getting upset, and your lack of feedback or an ETA is not 
> helping.
>
> -Joshua
>
> On Oct 6, 2020, at 11:47 AM, 'Olu' via Google App Engine <
> google-a...@googlegroups.com> wrote:
>
> Hi, Joshua
>
> Thank you for reporting this issue. It seems you may be affected by an 
> internal issue reported about App Engine standard applications returning 
> elev