You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@jclouds.apache.org by Sunil Shah <su...@mesosphere.io> on 2014/11/12 05:41:23 UTC
Nodes block forever on GCE
Hey User list,
We're seeing problems since upgrading to 2.0.0-SNAPSHOT from
1.8.2-SNAPSHOT. We're using jclouds-labs-google and when trying to
provision nodes, they block forever, despite the node now running:
2014-11-11 20:20:48,988 DEBUG [pool-3-thread-1] [] jclouds.compute
[SLF4JLogger.java:56] << RUNNING node(us-central1-a/test-cluster-6-626)
2014-11-11 20:20:49,001 DEBUG [pool-3-thread-3] [] jclouds.compute
[SLF4JLogger.java:56] >> blocking on sockets [107.178.210.31:22,
10.27.249.104:22] for 600000 MILLISECONDS
2014-11-11 20:20:52,044 DEBUG [pool-3-thread-2] [] jclouds.compute
[SLF4JLogger.java:56] << RUNNING node(us-central1-a/test-cluster-6-bec)
2014-11-11 20:20:52,045 DEBUG [pool-3-thread-8] [] jclouds.compute
[SLF4JLogger.java:56] >> blocking on sockets [146.148.92.89:22,
10.31.244.213:22] for 600000 MILLISECONDS
We haven't changed how we built the template:
builder.imageId(request.image)
.hardwareId(s"${request.zone}/${request.machineType}")
.locationId(request.zone)
.options(options.overrideLoginPrivateKey(Main.conf.privateKeyString))
.options(options.overrideLoginUser(request.sshUserName))
.options(options.authorizePublicKey(Main.conf.publicKeyString))
.build()
The serial output from the instance launched using 2.0.0-SNAPSHOT is here,
it looks like it can't find a boot disk:
https://gist.github.com/ssk2/7bfd47a1ee8965e51bdd
Serial output when the instance launches successfully using 1.8.2-SNAPSHOT
is here: https://gist.github.com/ssk2/19ae56c9f0d5f67cc5a5 using the same
image (backports-debian-7-wheezy-v20141108).
Any suggestions or ideas will be gratefully received :)
Sunil
Re: Nodes block forever on GCE
Posted by Sunil Shah <su...@mesosphere.io>.
Hey Ignasi - I'll have a look, thanks! We're
using backports-debian-7-wheezy-v20141108, which I had assumed would be
pretty safe. I'll try a few other images and see if that helps.
On Wed, Nov 12, 2014 at 12:21 AM, Ignasi Barrera <na...@apache.org> wrote:
> Looking at the image you're using, it seems to be a deprecated one (isn't
> it?).
>
> You may want to filter out the deprecated images and avoid using them (I
> recall having problems with disks when using some of them).
>
> In the current impl, the "deprecated" field, when present, is populated in
> the "userMetadata" map. You can safely filter out the images with the
> "deprecated" key. That can be done by using the "imageMatches" method from
> the TemplateBuilder with the appropriate predicate.
>
> HTH!
> El 12/11/2014 06:29, "Sunil Shah" <su...@mesosphere.io> escribió:
>
> Sounds good - thanks again for the information!
>>
>> On Tue, Nov 11, 2014 at 8:59 PM, Adrian Cole <ad...@gmail.com>
>> wrote:
>>
>>> We haven't discussed timeframes for 1.9, yet, Sunil.
>>>
>>> Danny, Ignasi and I have been working hard on GCE. Even if jclouds 1.9
>>> release isn't complete, GCE will likely be stable by the end of November.
>>>
>>> -A
>>> On Nov 11, 2014 8:51 PM, "Sunil Shah" <su...@mesosphere.io> wrote:
>>>
>>>> Aha, thanks for the information Adrian! Do you have an approximate ETA
>>>> for 1.9?
>>>>
>>>> Thanks,
>>>>
>>>> Sunil
>>>>
>>>> On Tue, Nov 11, 2014 at 8:48 PM, Adrian Cole <ad...@gmail.com>
>>>> wrote:
>>>>
>>>>> 2.0 snapshot has problems with firewall rules which prevent port 22
>>>>> from being reached. This will be fixed before that code is promoted as
>>>>> version 1.9.
>>>>>
>>>>> -A
>>>>> On Nov 11, 2014 8:42 PM, "Sunil Shah" <su...@mesosphere.io> wrote:
>>>>>
>>>>>> Hey User list,
>>>>>>
>>>>>> We're seeing problems since upgrading to 2.0.0-SNAPSHOT from
>>>>>> 1.8.2-SNAPSHOT. We're using jclouds-labs-google and when trying to
>>>>>> provision nodes, they block forever, despite the node now running:
>>>>>> 2014-11-11 20:20:48,988 DEBUG [pool-3-thread-1] [] jclouds.compute
>>>>>> [SLF4JLogger.java:56] << RUNNING node(us-central1-a/test-cluster-6-626)
>>>>>> 2014-11-11 20:20:49,001 DEBUG [pool-3-thread-3] [] jclouds.compute
>>>>>> [SLF4JLogger.java:56] >> blocking on sockets [107.178.210.31:22,
>>>>>> 10.27.249.104:22] for 600000 MILLISECONDS
>>>>>> 2014-11-11 20:20:52,044 DEBUG [pool-3-thread-2] [] jclouds.compute
>>>>>> [SLF4JLogger.java:56] << RUNNING node(us-central1-a/test-cluster-6-bec)
>>>>>> 2014-11-11 20:20:52,045 DEBUG [pool-3-thread-8] [] jclouds.compute
>>>>>> [SLF4JLogger.java:56] >> blocking on sockets [146.148.92.89:22,
>>>>>> 10.31.244.213:22] for 600000 MILLISECONDS
>>>>>>
>>>>>> We haven't changed how we built the template:
>>>>>> builder.imageId(request.image)
>>>>>> .hardwareId(s"${request.zone}/${request.machineType}")
>>>>>> .locationId(request.zone)
>>>>>>
>>>>>> .options(options.overrideLoginPrivateKey(Main.conf.privateKeyString))
>>>>>> .options(options.overrideLoginUser(request.sshUserName))
>>>>>> .options(options.authorizePublicKey(Main.conf.publicKeyString))
>>>>>> .build()
>>>>>>
>>>>>> The serial output from the instance launched using 2.0.0-SNAPSHOT is
>>>>>> here, it looks like it can't find a boot disk:
>>>>>> https://gist.github.com/ssk2/7bfd47a1ee8965e51bdd
>>>>>>
>>>>>> Serial output when the instance launches successfully using
>>>>>> 1.8.2-SNAPSHOT is here:
>>>>>> https://gist.github.com/ssk2/19ae56c9f0d5f67cc5a5 using the same
>>>>>> image (backports-debian-7-wheezy-v20141108).
>>>>>>
>>>>>> Any suggestions or ideas will be gratefully received :)
>>>>>>
>>>>>> Sunil
>>>>>>
>>>>>>
>>>>
>>
Re: Nodes block forever on GCE
Posted by Ignasi Barrera <na...@apache.org>.
Looking at the image you're using, it seems to be a deprecated one (isn't
it?).
You may want to filter out the deprecated images and avoid using them (I
recall having problems with disks when using some of them).
In the current impl, the "deprecated" field, when present, is populated in
the "userMetadata" map. You can safely filter out the images with the
"deprecated" key. That can be done by using the "imageMatches" method from
the TemplateBuilder with the appropriate predicate.
HTH!
El 12/11/2014 06:29, "Sunil Shah" <su...@mesosphere.io> escribió:
> Sounds good - thanks again for the information!
>
> On Tue, Nov 11, 2014 at 8:59 PM, Adrian Cole <ad...@gmail.com>
> wrote:
>
>> We haven't discussed timeframes for 1.9, yet, Sunil.
>>
>> Danny, Ignasi and I have been working hard on GCE. Even if jclouds 1.9
>> release isn't complete, GCE will likely be stable by the end of November.
>>
>> -A
>> On Nov 11, 2014 8:51 PM, "Sunil Shah" <su...@mesosphere.io> wrote:
>>
>>> Aha, thanks for the information Adrian! Do you have an approximate ETA
>>> for 1.9?
>>>
>>> Thanks,
>>>
>>> Sunil
>>>
>>> On Tue, Nov 11, 2014 at 8:48 PM, Adrian Cole <ad...@gmail.com>
>>> wrote:
>>>
>>>> 2.0 snapshot has problems with firewall rules which prevent port 22
>>>> from being reached. This will be fixed before that code is promoted as
>>>> version 1.9.
>>>>
>>>> -A
>>>> On Nov 11, 2014 8:42 PM, "Sunil Shah" <su...@mesosphere.io> wrote:
>>>>
>>>>> Hey User list,
>>>>>
>>>>> We're seeing problems since upgrading to 2.0.0-SNAPSHOT from
>>>>> 1.8.2-SNAPSHOT. We're using jclouds-labs-google and when trying to
>>>>> provision nodes, they block forever, despite the node now running:
>>>>> 2014-11-11 20:20:48,988 DEBUG [pool-3-thread-1] [] jclouds.compute
>>>>> [SLF4JLogger.java:56] << RUNNING node(us-central1-a/test-cluster-6-626)
>>>>> 2014-11-11 20:20:49,001 DEBUG [pool-3-thread-3] [] jclouds.compute
>>>>> [SLF4JLogger.java:56] >> blocking on sockets [107.178.210.31:22,
>>>>> 10.27.249.104:22] for 600000 MILLISECONDS
>>>>> 2014-11-11 20:20:52,044 DEBUG [pool-3-thread-2] [] jclouds.compute
>>>>> [SLF4JLogger.java:56] << RUNNING node(us-central1-a/test-cluster-6-bec)
>>>>> 2014-11-11 20:20:52,045 DEBUG [pool-3-thread-8] [] jclouds.compute
>>>>> [SLF4JLogger.java:56] >> blocking on sockets [146.148.92.89:22,
>>>>> 10.31.244.213:22] for 600000 MILLISECONDS
>>>>>
>>>>> We haven't changed how we built the template:
>>>>> builder.imageId(request.image)
>>>>> .hardwareId(s"${request.zone}/${request.machineType}")
>>>>> .locationId(request.zone)
>>>>>
>>>>> .options(options.overrideLoginPrivateKey(Main.conf.privateKeyString))
>>>>> .options(options.overrideLoginUser(request.sshUserName))
>>>>> .options(options.authorizePublicKey(Main.conf.publicKeyString))
>>>>> .build()
>>>>>
>>>>> The serial output from the instance launched using 2.0.0-SNAPSHOT is
>>>>> here, it looks like it can't find a boot disk:
>>>>> https://gist.github.com/ssk2/7bfd47a1ee8965e51bdd
>>>>>
>>>>> Serial output when the instance launches successfully using
>>>>> 1.8.2-SNAPSHOT is here:
>>>>> https://gist.github.com/ssk2/19ae56c9f0d5f67cc5a5 using the same
>>>>> image (backports-debian-7-wheezy-v20141108).
>>>>>
>>>>> Any suggestions or ideas will be gratefully received :)
>>>>>
>>>>> Sunil
>>>>>
>>>>>
>>>
>
Re: Nodes block forever on GCE
Posted by Sunil Shah <su...@mesosphere.io>.
Sounds good - thanks again for the information!
On Tue, Nov 11, 2014 at 8:59 PM, Adrian Cole <ad...@gmail.com>
wrote:
> We haven't discussed timeframes for 1.9, yet, Sunil.
>
> Danny, Ignasi and I have been working hard on GCE. Even if jclouds 1.9
> release isn't complete, GCE will likely be stable by the end of November.
>
> -A
> On Nov 11, 2014 8:51 PM, "Sunil Shah" <su...@mesosphere.io> wrote:
>
>> Aha, thanks for the information Adrian! Do you have an approximate ETA
>> for 1.9?
>>
>> Thanks,
>>
>> Sunil
>>
>> On Tue, Nov 11, 2014 at 8:48 PM, Adrian Cole <ad...@gmail.com>
>> wrote:
>>
>>> 2.0 snapshot has problems with firewall rules which prevent port 22 from
>>> being reached. This will be fixed before that code is promoted as version
>>> 1.9.
>>>
>>> -A
>>> On Nov 11, 2014 8:42 PM, "Sunil Shah" <su...@mesosphere.io> wrote:
>>>
>>>> Hey User list,
>>>>
>>>> We're seeing problems since upgrading to 2.0.0-SNAPSHOT from
>>>> 1.8.2-SNAPSHOT. We're using jclouds-labs-google and when trying to
>>>> provision nodes, they block forever, despite the node now running:
>>>> 2014-11-11 20:20:48,988 DEBUG [pool-3-thread-1] [] jclouds.compute
>>>> [SLF4JLogger.java:56] << RUNNING node(us-central1-a/test-cluster-6-626)
>>>> 2014-11-11 20:20:49,001 DEBUG [pool-3-thread-3] [] jclouds.compute
>>>> [SLF4JLogger.java:56] >> blocking on sockets [107.178.210.31:22,
>>>> 10.27.249.104:22] for 600000 MILLISECONDS
>>>> 2014-11-11 20:20:52,044 DEBUG [pool-3-thread-2] [] jclouds.compute
>>>> [SLF4JLogger.java:56] << RUNNING node(us-central1-a/test-cluster-6-bec)
>>>> 2014-11-11 20:20:52,045 DEBUG [pool-3-thread-8] [] jclouds.compute
>>>> [SLF4JLogger.java:56] >> blocking on sockets [146.148.92.89:22,
>>>> 10.31.244.213:22] for 600000 MILLISECONDS
>>>>
>>>> We haven't changed how we built the template:
>>>> builder.imageId(request.image)
>>>> .hardwareId(s"${request.zone}/${request.machineType}")
>>>> .locationId(request.zone)
>>>>
>>>> .options(options.overrideLoginPrivateKey(Main.conf.privateKeyString))
>>>> .options(options.overrideLoginUser(request.sshUserName))
>>>> .options(options.authorizePublicKey(Main.conf.publicKeyString))
>>>> .build()
>>>>
>>>> The serial output from the instance launched using 2.0.0-SNAPSHOT is
>>>> here, it looks like it can't find a boot disk:
>>>> https://gist.github.com/ssk2/7bfd47a1ee8965e51bdd
>>>>
>>>> Serial output when the instance launches successfully using
>>>> 1.8.2-SNAPSHOT is here:
>>>> https://gist.github.com/ssk2/19ae56c9f0d5f67cc5a5 using the same image
>>>> (backports-debian-7-wheezy-v20141108).
>>>>
>>>> Any suggestions or ideas will be gratefully received :)
>>>>
>>>> Sunil
>>>>
>>>>
>>
Re: Nodes block forever on GCE
Posted by Adrian Cole <ad...@gmail.com>.
We haven't discussed timeframes for 1.9, yet, Sunil.
Danny, Ignasi and I have been working hard on GCE. Even if jclouds 1.9
release isn't complete, GCE will likely be stable by the end of November.
-A
On Nov 11, 2014 8:51 PM, "Sunil Shah" <su...@mesosphere.io> wrote:
> Aha, thanks for the information Adrian! Do you have an approximate ETA for
> 1.9?
>
> Thanks,
>
> Sunil
>
> On Tue, Nov 11, 2014 at 8:48 PM, Adrian Cole <ad...@gmail.com>
> wrote:
>
>> 2.0 snapshot has problems with firewall rules which prevent port 22 from
>> being reached. This will be fixed before that code is promoted as version
>> 1.9.
>>
>> -A
>> On Nov 11, 2014 8:42 PM, "Sunil Shah" <su...@mesosphere.io> wrote:
>>
>>> Hey User list,
>>>
>>> We're seeing problems since upgrading to 2.0.0-SNAPSHOT from
>>> 1.8.2-SNAPSHOT. We're using jclouds-labs-google and when trying to
>>> provision nodes, they block forever, despite the node now running:
>>> 2014-11-11 20:20:48,988 DEBUG [pool-3-thread-1] [] jclouds.compute
>>> [SLF4JLogger.java:56] << RUNNING node(us-central1-a/test-cluster-6-626)
>>> 2014-11-11 20:20:49,001 DEBUG [pool-3-thread-3] [] jclouds.compute
>>> [SLF4JLogger.java:56] >> blocking on sockets [107.178.210.31:22,
>>> 10.27.249.104:22] for 600000 MILLISECONDS
>>> 2014-11-11 20:20:52,044 DEBUG [pool-3-thread-2] [] jclouds.compute
>>> [SLF4JLogger.java:56] << RUNNING node(us-central1-a/test-cluster-6-bec)
>>> 2014-11-11 20:20:52,045 DEBUG [pool-3-thread-8] [] jclouds.compute
>>> [SLF4JLogger.java:56] >> blocking on sockets [146.148.92.89:22,
>>> 10.31.244.213:22] for 600000 MILLISECONDS
>>>
>>> We haven't changed how we built the template:
>>> builder.imageId(request.image)
>>> .hardwareId(s"${request.zone}/${request.machineType}")
>>> .locationId(request.zone)
>>>
>>> .options(options.overrideLoginPrivateKey(Main.conf.privateKeyString))
>>> .options(options.overrideLoginUser(request.sshUserName))
>>> .options(options.authorizePublicKey(Main.conf.publicKeyString))
>>> .build()
>>>
>>> The serial output from the instance launched using 2.0.0-SNAPSHOT is
>>> here, it looks like it can't find a boot disk:
>>> https://gist.github.com/ssk2/7bfd47a1ee8965e51bdd
>>>
>>> Serial output when the instance launches successfully using
>>> 1.8.2-SNAPSHOT is here:
>>> https://gist.github.com/ssk2/19ae56c9f0d5f67cc5a5 using the same image
>>> (backports-debian-7-wheezy-v20141108).
>>>
>>> Any suggestions or ideas will be gratefully received :)
>>>
>>> Sunil
>>>
>>>
>
Re: Nodes block forever on GCE
Posted by Sunil Shah <su...@mesosphere.io>.
Aha, thanks for the information Adrian! Do you have an approximate ETA for
1.9?
Thanks,
Sunil
On Tue, Nov 11, 2014 at 8:48 PM, Adrian Cole <ad...@gmail.com>
wrote:
> 2.0 snapshot has problems with firewall rules which prevent port 22 from
> being reached. This will be fixed before that code is promoted as version
> 1.9.
>
> -A
> On Nov 11, 2014 8:42 PM, "Sunil Shah" <su...@mesosphere.io> wrote:
>
>> Hey User list,
>>
>> We're seeing problems since upgrading to 2.0.0-SNAPSHOT from
>> 1.8.2-SNAPSHOT. We're using jclouds-labs-google and when trying to
>> provision nodes, they block forever, despite the node now running:
>> 2014-11-11 20:20:48,988 DEBUG [pool-3-thread-1] [] jclouds.compute
>> [SLF4JLogger.java:56] << RUNNING node(us-central1-a/test-cluster-6-626)
>> 2014-11-11 20:20:49,001 DEBUG [pool-3-thread-3] [] jclouds.compute
>> [SLF4JLogger.java:56] >> blocking on sockets [107.178.210.31:22,
>> 10.27.249.104:22] for 600000 MILLISECONDS
>> 2014-11-11 20:20:52,044 DEBUG [pool-3-thread-2] [] jclouds.compute
>> [SLF4JLogger.java:56] << RUNNING node(us-central1-a/test-cluster-6-bec)
>> 2014-11-11 20:20:52,045 DEBUG [pool-3-thread-8] [] jclouds.compute
>> [SLF4JLogger.java:56] >> blocking on sockets [146.148.92.89:22,
>> 10.31.244.213:22] for 600000 MILLISECONDS
>>
>> We haven't changed how we built the template:
>> builder.imageId(request.image)
>> .hardwareId(s"${request.zone}/${request.machineType}")
>> .locationId(request.zone)
>>
>> .options(options.overrideLoginPrivateKey(Main.conf.privateKeyString))
>> .options(options.overrideLoginUser(request.sshUserName))
>> .options(options.authorizePublicKey(Main.conf.publicKeyString))
>> .build()
>>
>> The serial output from the instance launched using 2.0.0-SNAPSHOT is
>> here, it looks like it can't find a boot disk:
>> https://gist.github.com/ssk2/7bfd47a1ee8965e51bdd
>>
>> Serial output when the instance launches successfully using
>> 1.8.2-SNAPSHOT is here: https://gist.github.com/ssk2/19ae56c9f0d5f67cc5a5
>> using the same image (backports-debian-7-wheezy-v20141108).
>>
>> Any suggestions or ideas will be gratefully received :)
>>
>> Sunil
>>
>>
Re: Nodes block forever on GCE
Posted by Adrian Cole <ad...@gmail.com>.
2.0 snapshot has problems with firewall rules which prevent port 22 from
being reached. This will be fixed before that code is promoted as version
1.9.
-A
On Nov 11, 2014 8:42 PM, "Sunil Shah" <su...@mesosphere.io> wrote:
> Hey User list,
>
> We're seeing problems since upgrading to 2.0.0-SNAPSHOT from
> 1.8.2-SNAPSHOT. We're using jclouds-labs-google and when trying to
> provision nodes, they block forever, despite the node now running:
> 2014-11-11 20:20:48,988 DEBUG [pool-3-thread-1] [] jclouds.compute
> [SLF4JLogger.java:56] << RUNNING node(us-central1-a/test-cluster-6-626)
> 2014-11-11 20:20:49,001 DEBUG [pool-3-thread-3] [] jclouds.compute
> [SLF4JLogger.java:56] >> blocking on sockets [107.178.210.31:22,
> 10.27.249.104:22] for 600000 MILLISECONDS
> 2014-11-11 20:20:52,044 DEBUG [pool-3-thread-2] [] jclouds.compute
> [SLF4JLogger.java:56] << RUNNING node(us-central1-a/test-cluster-6-bec)
> 2014-11-11 20:20:52,045 DEBUG [pool-3-thread-8] [] jclouds.compute
> [SLF4JLogger.java:56] >> blocking on sockets [146.148.92.89:22,
> 10.31.244.213:22] for 600000 MILLISECONDS
>
> We haven't changed how we built the template:
> builder.imageId(request.image)
> .hardwareId(s"${request.zone}/${request.machineType}")
> .locationId(request.zone)
> .options(options.overrideLoginPrivateKey(Main.conf.privateKeyString))
> .options(options.overrideLoginUser(request.sshUserName))
> .options(options.authorizePublicKey(Main.conf.publicKeyString))
> .build()
>
> The serial output from the instance launched using 2.0.0-SNAPSHOT is here,
> it looks like it can't find a boot disk:
> https://gist.github.com/ssk2/7bfd47a1ee8965e51bdd
>
> Serial output when the instance launches successfully using 1.8.2-SNAPSHOT
> is here: https://gist.github.com/ssk2/19ae56c9f0d5f67cc5a5 using the same
> image (backports-debian-7-wheezy-v20141108).
>
> Any suggestions or ideas will be gratefully received :)
>
> Sunil
>
>