GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/13
removing git pull unsupported line from README
Testing pull requests!
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/navina/samza test-branch
Github user navina closed the pull request at:
https://github.com/apache/samza/pull/13
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/16
SAMZA-1030 : Add documentation for change in the contribution process
I re-organized some of the website files and created a "Contributor's
Corner" that collates all info related to n
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/17
SAMZA-1033: Remove import-control from checkstyle
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/navina/samza SAMZA-1033
Alternatively you can
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/27
Adding Build Status to README.md
I setup a Jenkins Build Job to get triggered every time changes are pushed
to samza repository.
https://builds.apache.org/view/S-Z/view/Samza/job/samza-freestyle
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/44
Initial Standalone JobCoordinator and StreamProcessor API
This patch contains changes associated with the Standalone StreamProcessor,
where there is no coordination. This will work for load-balanced
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/48
SAMZA-1082 : Implement Leader Election using ZK
Simple implementation of leader election recipe along with unit tests
You can merge this pull request into a Git repository by running:
$ git pull
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/54
SAMZA-1084 - User thread does not see errors from the processor thread
Adding a ProcessorLifecycleCallback for specifying user-defined error
handling
You can merge this pull request into a Git
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/102
SAMZA-1175 - Removing CoordinationService from JobCoordinatorFactory
interface
@sborya I have made the changes. Please review. Thanks!
You can merge this pull request into a Git repository by
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/103
SAMZA-1126 - Semantics of processorId in Samza
Implementation based on
[SEP-1](https://cwiki.apache.org/confluence/display/SAMZA/SEP-1%3A+Semantics+of+ProcessorId+in+Samza)
You can merge this
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/107
SAMZA-1182 - Commenting out some of the flaky tests
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/navina/samza SAMZA-1182
Alternatively you
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/112
Samza-1187 : TestZkProcessorLatch tests share state causing transient
failures in CI builds
* Removed testSingleCountdown - didn't understand what was being tested
* The timeout behavio
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/121
SAMZA-1208 - IllegalFormatConversionException in LocalContainerRunner
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/navina/samza SAMZA-1208
Github user navina closed the pull request at:
https://github.com/apache/samza/pull/121
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is
GitHub user navina reopened a pull request:
https://github.com/apache/samza/pull/121
SAMZA-1208 - IllegalFormatConversionException in LocalContainerRunner
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/navina/samza SAMZA-1208
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/125
SAMZA-1213 - StreamProcessorLifeCycleAware interface should not use
processorId
Refactoring LocalApplicationRunner s.t. each processor has its own listener
instance, instead of a single listener
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/139
SAMZA-1220 : Add thread name to SamzaContainer shutdown hook and prevent
shutdown deadlock
* SamzaContainerExceptionHandler is written in Java and used by
LocalContainerRunner.java
You can merge
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/146
SAMZA-1224 : Revert job coordinator factory config to the old format
We didn't release since adding this config. So, it is ok to change the
format now.
You can merge this pull request into
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/148
SAMZA-1212 - Refactor interaction between StreamProcessor, JobCoordinator
and SamzaContainer
See SAMZA-1212 for motivation toward this refactoring.
Changes here are:
* Removed awaitStart
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/153
SAMZA-1251 - Remove DebounceTimer dependency from ZkLeaderElector
Fyi: This PR depends on PR #148
You can merge this pull request into a Git repository by running:
$ git pull https://github.com
Github user navina closed the pull request at:
https://github.com/apache/samza/pull/54
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/162
SAMZA-1228 : StreamProcessor should stop JmxServer
This is not the solution posted in SAMZA-1228. For now, we are moving
jmxserver lifecycle to be within the container. Ideally, it should be within
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/166
SAMZA-1150 : Handling Error propagation between ZkJobCoordinator &
DebounceTimer
* Treats all errors in jobcoordinator as FATAL and shuts-down the
streamprocessor
* [Bug] Fixed bug reporte
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/173
SAMZA-1272 : ZkCoordinationUtils deletes the entire Zk tree on reset
* ZkCoordinationUtils has a reset interface that deletes the entire Zk
tree. This is not desirable.
* Also, fixed flakiness in
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/179
SAMZA-1276 : Adding customReporters passed by StreamProcessor into
SamzaContainer
SamzaContainer is not adding custom reporters along side class-loaded
reporters from config. This was missed in
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/195
SAMZA-1128 : Remove dependency of debounce timer from the CoordinationUtils
This patch addresses the following:
* Removes CoordinationUtils#getBarrier, BarrierForVersionUpgrade interface
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/201
Updating committer page on samza.apache.org
Giving more screen time for currently active Apache Samza committers (no
JIRA)
You can merge this pull request into a Git repository by running
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/202
SAMZA-1307 - Fix ZkKeyBuilder null checks for pathPrefix
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/navina/samza SAMZA-1307
Alternatively
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/230
SAMZA-1340 - StreamProcessor does not propagate container failures from
StreamTask
Storing the exception seen from the container in the
`SamzaContainerListener#onFailure(Throwable)` in the
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/274
SAMZA-1396 TestZkLocalApplicationRunner tests fails after SAMZA-1385
* Fixes ZkPath issues
* Fixes appname / jobname mismatch
You can merge this pull request into a Git repository by running
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/301
Samza versioned Release Notes
Adding a versioned page for release/upgrade notes. We can start this
process from the next major version release, aka 0.14.0.
Please update this page as and
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/302
Fixing broken link between Yarn Host Affinity and Resource Localizatiâ¦
Fixing broken link between Yarn Host Affinity and Resource Localization
pages under Documentation
Patch needs to be
GitHub user navina opened a pull request:
https://github.com/apache/samza/pull/379
SAMZA-1523 Cleanup table entries before shutting down the processor
Modified the `TableUtils#deleteProcessorEntity` to provide an option to
disable optimistic locking during a call to Azure Table
Yep!
On 1/23/15, 9:56 AM, "Yi Pan" wrote:
>Yes!
>
>On Fri, Jan 23, 2015 at 9:49 AM, Chris Riccomini
>wrote:
>
>> Hey all,
>>
>> Could you please confirm that you're seeing this? I'm trying to verify
>>the
>> TLP migration for:
>>
>> https://issues.apache.org/jira/browse/INFRA-9055
>>
>> Cheer
ile831263line26>
> >
> > Would org.apache.kafka.clients.producer.MockProducer be configured to
> > satisfy the requests here?
I have addresssed the same question in
https://reviews.apache.org/r/29899/#comment113812
- Navina
-
ni wrote:
> > samza-test/src/main/config/negate-number.properties, line 47
> > <https://reviews.apache.org/r/29899/diff/4/?file=831268#file831268line47>
> >
> > Can you do a quick grep through the code base, adn make
clean build - SUCCESSFUL
Thanks,
Navina Ramesh
.kafka.clients.producer.MockProducer be configured to
> > satisfy the requests here?
>
> Navina Ramesh wrote:
> I have addresssed the same question in
> https://reviews.apache.org/r/29899/#comment113812
Created a follow-up task here - ht
negate-number.properties
4989b279a22ffd8acdb2c31ca4813f6768edc75b
samza-test/src/test/scala/org/apache/samza/test/integration/TestStatefulTask.scala
ca25258217e5ebc44b34fbc4d69ecb28c81df618
Diff: https://reviews.apache.org/r/29899/diff/
Testing
---
./gradlew clean build - SUCCESSFUL
Thanks,
Navina Ramesh
/java/org/apache/samza/task/sql/StreamSqlTask.java
PRE-CREATION
settings.gradle 3a01fd66359b8c79954ae8f34eeaf4b2e3fdc0b4
Diff: https://reviews.apache.org/r/30634/diff/
Testing
---
Thanks,
Navina Ramesh
://reviews.apache.org/r/30627/ . Main changes are: SqlStringSerde,
SqlStringSerdeFactory, StringData and StringSchema
- Navina Ramesh
On Feb. 4, 2015, 7:39 p.m., Navina Ramesh wrote:
>
> ---
> This is an automatically generated e-mail.
I have updates SAMZA-484 and RB is here -
https://reviews.apache.org/r/30634/diff/#
Navina
On 2/4/15, 11:13 AM, "Yi Pan" wrote:
>Just did the update w/ SAMZA-482.
>
>On Wed, Feb 4, 2015 at 10:31 AM, Chris Riccomini
>wrote:
>
>> I think so. There was some RB down
ine 141). Anything with "serializers.registry."
substring is parsed and used to generate a map of form "serdeType =>
serdeFactory instance". This is used by SerdeManager for lookups.
For example,
if a job defines
"serializers.registry.string.class=org.apache.samza.se
hanks,
Navina Ramesh
ATION
settings.gradle bb07a3b84b14dcef94da1bb166eab6aa3d0026bb
Diff: https://reviews.apache.org/r/30634/diff/
Testing
---
Thanks,
Navina Ramesh
tps://reviews.apache.org/r/30781/#comment117547>
What are the final job states that is returned by Yarn?
samza-test/src/main/python/tests/performance_tests.py
<https://reviews.apache.org/r/30781/#comment117545>
What is the purpose of this if statement ?
- Navina Ramesh
On Feb. 9
+1 on using the right terminology.
Navina
On 2/17/15, 9:33 AM, "Yi Pan" wrote:
>+1 on consolidating the terminology as well. Azure's definition looks good
>to me.
>
>On Tue, Feb 17, 2015 at 9:13 AM, Chris Riccomini
>wrote:
>
>> Hey Julian,
>>
>
).
Checkpointing is synchronous, if you use KafkaCheckpointManager. In that
implementation, you can take a look at the ³writeLog² method that makes a
synchronous call to the KafkaProducer.
Hope this helps!
Cheers,
Navina
On 2/19/15, 3:59 PM, "Richard Lee" wrote:
>
>
>> On Feb 19, 201
---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/31405/#review74065
---
Looks good!
- Navina Ramesh
On Feb. 25, 2015, 5:54 a.m., Milinda
/
Testing
---
Thanks,
Navina Ramesh
;t you do the failedSources empty check here instead of Line 153 ?
- Navina Ramesh
On March 17, 2015, 9:37 a.m., Yan Fang wrote:
>
> ---
> This is an automatically generated e-mail. To reply, visit:
://reviews.apache.org/r/32188/diff/
Testing (updated)
---
./bin/check-all.sh ran successfully on MAC
Thanks,
Navina Ramesh
> > nit: do we need multiple instances of ByteSerde here?
Ah..not really. Will change that.
- Navina
---
This is an automatically generated e-mail. To reply, visit:
https://
Diff: https://reviews.apache.org/r/32188/diff/
Testing
---
./bin/check-all.sh ran successfully on MAC
Thanks,
Navina Ramesh
---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/32155/#review77293
---
Ship it!
Ship It!
- Navina Ramesh
On March 20, 2015, 9:40 p.m
org.apache.samza.container.TaskInstance$$anonfun$process$1.apply$mcV$sp(Tas
kInstance.scala:133)
Did you make the changes to fix this error? Your patch doesn¹t seem to
have that.
Line 38 Map jsonObject = (Map)
envelope.getMessage();
Lmk so I can investigate further.
Cheers!
Navina
On 3/23/15, 6:43 PM, "Ash W Mat
s a reputation of 1500 and my
reputation is just 1 :)
If any stackoverflow enthusiast here wants to earn some good karma, please help
us in creating a tag - “apache-samza” in stack overflow.
Thanks!
Navina
Did you remove those lines as well?
Navina
On 3/23/15, 8:31 PM, "Ash W Matheson" wrote:
>Just looking at the diff I posted and it's:
>
>
> 1. try {
> 2. - Map parsedJsonObject =
>parse(event.getRawEvent(
> ));
> 3. + S
a new Samza user, please let us know if it is hard to find
any information.
Thanks!
Navina
On 3/24/15, 12:38 AM, "Jordi Blasi Uribarri" wrote:
>I tried to do it in the question you (kindly) answered and found that
>same problem. I only have 150 reputation (I have some d
://samza.apache.org/learn/documentation/latest/jobs/configuration-table.html)
with info about this re-writer ?
- Navina Ramesh
On April 17, 2015, 1:12 p.m., Dan Harvey wrote:
>
> ---
> This is an automatically generated e-mail. To reply, visit
/elasticsearch/ElasticsearchSystemProducer.java
<https://reviews.apache.org/r/33297/#comment131450>
Is this flush a blocking call in elasticsearch?
- Navina Ramesh
On April 22, 2015, 5:12 p.m., Dan Harvey wrote:
>
> ---
>
> > If the previous send has failed, i.e., sendFailed is true here,
> > shouldn't we throw an exception here ? Otherwise it might result in message
> > develivered out of order.
I think the messages are only buffered and not sent until we i
/RocksDbKeyValueStore.scala
66c2a0dc2e38e21f951727a30f0987776ac52fe2
Diff: https://reviews.apache.org/r/33453/diff/
Testing
---
Thanks,
Navina Ramesh
.
Note: you have to set an environment variable LOGGED_STORE_BASE_DIR pointing to
the new location to persist the changelog attached stores. Otherwise, it will
default to YARN's cwd and will not re-use local state.
Thanks,
Navina Ramesh
environment variable LOGGED_STORE_BASE_DIR pointing to
the new location to persist the changelog attached stores. Otherwise, it will
default to YARN's cwd and will not re-use local state.
Thanks,
Navina Ramesh
s here?
- Navina Ramesh
On April 23, 2015, 6:18 p.m., Dan Harvey wrote:
>
> ---
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.
> >
> > From the store initiation code, it seems that storeBaseDir won't have
> > any logged store paths if logged store base dir is configured. Hence, here
> > we may be deleting empty/non-existing storagePar
Question: shouldn't the base dir to be determined by container ID, not
> > jobId?
>
> Navina Ramesh wrote:
> I think containerId is not really relevant because the store partition
> directories are uniquely named across the containers. The ownership of
> containers
hanks,
Navina Ramesh
s there any checksum in the offset file to make sure that we don't
> > read in a corrupted value?
>
> Navina Ramesh wrote:
> No. As of now, we don't have any checksum. Do you think it is better to
> add it now than later ?
Yi,
I created SAMZA
---
Tested locally using hello-samza.
Note: you have to set an environment variable LOGGED_STORE_BASE_DIR pointing to
the new location to persist the changelog attached stores. Otherwise, it will
default to YARN's cwd and will not re-use local state.
Thanks,
Navina Ramesh
/org/apache/samza/util/Util.scala, line 274
> > <https://reviews.apache.org/r/33453/diff/4/?file=950517#file950517line274>
> >
> > nit: log ("CRC doesn't match, data corrupted, skipping file") or
> > something along those lines.
Ok. Makes sense.
- Navina
r changelogged over logged, as it seems consistent with other
> > usages in our code.
I don't see any "changelogged" usage. It actually makes the variable name
really long.
Since loggedStore can't really mean anything other than a
za/container/SamzaContainer.scala,
> > lines 508-509
> > <https://reviews.apache.org/r/33453/diff/4/?file=950515#file950515line508>
> >
> > why do we need 2 storeDir? Is it because we want to delete the default
> > state Dir?
Yes
> On
locally using hello-samza.
Note: you have to set an environment variable LOGGED_STORE_BASE_DIR pointing to
the new location to persist the changelog attached stores. Otherwise, it will
default to YARN's cwd and will not re-use local state.
Thanks,
Navina Ramesh
ould this logic be another method, such as
> > readOffsetFromFile, not "cleanBaseDir"? A little misleading.
>
> Navina Ramesh wrote:
> I can modularize it further if it helps. Since we need to load offsets
> (if available) before cleaning directories, it became a p
ngeLogSystemStreamPartition matter here?
> > 2. prefer the storePartitionDir, because it's the dir for each
> > partition/task, right?
>
> Navina Ramesh wrote:
> 1. why does the changeLogSystemStreamPartition matter here?
> >> changeLogSystemStreamParti
sk/sql/RandomWindowOperatorTask.java
<https://reviews.apache.org/r/34009/#comment134678>
Shouldn't the window id be system generated?
- Navina Ramesh
On May 13, 2015, 5:36 p.m., Yi Pan (Data Infrastructure) wrote:
>
> ---
> This is an automati
> On May 13, 2015, 9:56 p.m., Navina Ramesh wrote:
> > samza-sql-core/src/test/java/org/apache/samza/task/sql/RandomWindowOperatorTask.java,
> > line 96
> > <https://reviews.apache.org/r/34009/diff/1/?file=954320#file954320line96>
> >
> > Shouldn'
Hi Shekar,
Actually, I am not able to view even after signing in. Like Yan suggested,
please try to share it elsewhere on the internet. Should G drive work?
Thanks!
Navina
On 5/21/15, 11:21 AM, "Yan Fang" wrote:
>Hi Shekar,
>
>This website requires signed in. Could you p
names, irrespective of whether it is storename or taskname. Just my two cents.
- Navina Ramesh
On May 20, 2015, 10:04 p.m., Yan Fang wrote:
>
> ---
> This is an automaticall
g up YARN on 3
machines.
Verified that the message is correctly written and consumed from the
Coordinator Stream
Thanks,
Navina Ramesh
omment137337>
Why do we need 2 instances of the TopologyBuilder here?
I think this occurs because stream() and sink() method return
OperatorSource type rather than a TopologyBuilder instance.
How differently does the TopologyBuilder handle the OperatorSource and the
OperatorS
ttps://reviews.apache.org/r/34746/diff/
Testing
---
Used a sample job to test it locally and also, by setting up YARN on 3
machines.
Verified that the message is correctly written and consumed from the
Coordinator Stream
Thanks,
Navina Ramesh
> On May 29, 2015, 1:28 a.m., Navina Ramesh wrote:
> > samza-sql-core/src/test/java/org/apache/samza/task/sql/UserCallbacksSqlTask.java,
> > line 120
> > <https://reviews.apache.org/r/34500/diff/1/?file=965740#file965740line120>
> >
> > I thought T
73286#file973286line629>
> >
> > catch the UnknownHostException ?
Mmm.. makes sense. It could be thrown. I will change it to handle the exception
with a no-op.
- Navina
---
This is an automatically generated e-m
Hi Zach,
It sounds like you are asking for a SystemConsumer for hdfs. Does
SAMZA-263 match your requirements?
Thanks!
Navina
On 5/29/15, 2:23 PM, "Zach Cox" wrote:
>(continuing from previous email) in addition to not wanting to duplicate
>code, say that some of the Samza jobs
messages from the other Kafka topic.
Thanks!
Navina
On 5/29/15, 2:26 PM, "Navina Ramesh" wrote:
>Hi Zach,
>
>It sounds like you are asking for a SystemConsumer for hdfs. Does
>SAMZA-263 match your requirements?
>
>Thanks!
>Navina
>
>On 5/29/15, 2:23 PM, &q
Hi Zach,
Regarding the JIRA, it is assigned to Jakob Homan. He will be the right
person to comment on that.
Thanks!
Navina
On 5/29/15, 2:33 PM, "Zach Cox" wrote:
>Hi Navina,
>
>I did see that jira and it would definitely be useful. I was thinking of
>maybe trying to bu
would be to consume/bootstrap directly from HDFS :)
Cheers!
Navina
On 5/29/15, 2:44 PM, "Zach Cox" wrote:
>Hi Navina,
>
>A similar approach I considered was using an infinite/very large retention
>period on those event kafka topics, so they would always contain all
&g
streams.
Cheers!
Navina
[1]
https://github.com/apache/samza/blob/master/samza-api/src/main/java/org/apa
che/samza/system/chooser/MessageChooser.java#L26
On 5/29/15, 3:07 PM, "Zach Cox" wrote:
>Hi Navina,
>
>Do you mean bootstrapping from hdfs as in [1]? That is an interes
Diff: https://reviews.apache.org/r/34746/diff/
Testing
---
Used a sample job to test it locally and also, by setting up YARN on 3
machines.
Verified that the message is correctly written and consumed from the
Coordinator Stream
Thanks,
Navina Ramesh
a
> > single CoordinatorStreamManager? And
> > CoordinatorStreamManager.getCheckpointManager()/getChangelogManager()/getLocalityManager()
> > would return the specific management function handler?
I guess this is also a pattern we can follow during the refactoring.
As men
Diff: https://reviews.apache.org/r/34746/diff/
Testing
---
Used a sample job to test it locally and also, by setting up YARN on 3
machines.
Verified that the message is correctly written and consumed from the
Coordinator Stream
Thanks,
Navina Ramesh
Milinda Pathirage
>wrote:
>
>> Hi Navina,
>>
>> Did we decided to push this patch to samza-sql branch. I thought Yi is
>> still working on this. Some Git conflict related texts are still there
>>in
>> this commit.
>>
>> +<<<<<<
rg/apache/samza/config/KafkaConfig.scala
<https://reviews.apache.org/r/35067/#comment139191>
We need a log message when this value is overriden (to < INT.MAX) and that
Samza cannot handle producer failures. Hence, the onus of retries is left up to
the kafka producer.
- Navina Ramesh
While Util.envVarEscape() is not exactly smart in the escaping characters,
it is a vital minimum to have.
Thanks!
Navina
On 6/9/15, 5:22 AM, "Tommy Becker" wrote:
>Ok, the issue with spaces appears to have been addressed by
>https://issues.apache.org/jira/browse/SAMZA-530. But quo
Hi Mart,
Tommy is right. We moved the serializers into samza-core. I think that
documentation requires an update. Please remove the dependency section for
samza_serializers.
Thanks!
Navina
On 6/10/15, 4:36 AM, "Tommy Becker" wrote:
>I'm sure others will chime in, but I
Hey Mart,
I created a task for this -> https://issues.apache.org/jira/browse/SAMZA-707
I will try to update it today.
Thanks!
Navina
From: Mart Haitjema
mailto:mhaitj...@skyportsystems.com>>
Reply-To: mailto:dev@samza.apache.org>>
Date: Wednesday, June 10, 2015 at 1:36 PM
---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/35071/#review87507
---
Ship it!
lgtm!
- Navina Ramesh
On June 4, 2015, 2:42 p.m
1 - 100 of 596 matches
Mail list logo
