[jira] [Commented] (MESOS-3771) Mesos JSON API creates invalid JSON due to lack of binary data / non-ASCII handling
[ https://issues.apache.org/jira/browse/MESOS-3771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14970102#comment-14970102 ] Joseph Wu commented on MESOS-3771: -- Created [MESOS-3794] to track the point #2 above. That portion will likely be a more involved change. > Mesos JSON API creates invalid JSON due to lack of binary data / non-ASCII > handling > --- > > Key: MESOS-3771 > URL: https://issues.apache.org/jira/browse/MESOS-3771 > Project: Mesos > Issue Type: Bug > Components: HTTP API >Affects Versions: 0.24.1, 0.26.0 >Reporter: Steven Schlansker >Assignee: Joseph Wu >Priority: Critical > Labels: mesosphere > > Spark encodes some binary data into the ExecutorInfo.data field. This field > is sent as a "bytes" Protobuf value, which can have arbitrary non-UTF8 data. > If you have such a field, it seems that it is splatted out into JSON without > any regards to proper character encoding: > {code} > 0006b0b0 2e 73 70 61 72 6b 2e 65 78 65 63 75 74 6f 72 2e |.spark.executor.| > 0006b0c0 4d 65 73 6f 73 45 78 65 63 75 74 6f 72 42 61 63 |MesosExecutorBac| > 0006b0d0 6b 65 6e 64 22 7d 2c 22 64 61 74 61 22 3a 22 ac |kend"},"data":".| > 0006b0e0 ed 5c 75 30 30 30 30 5c 75 30 30 30 35 75 72 5c |.\u\u0005ur\| > 0006b0f0 75 30 30 30 30 5c 75 30 30 30 66 5b 4c 73 63 61 |u\u000f[Lsca| > 0006b100 6c 61 2e 54 75 70 6c 65 32 3b 2e cc 5c 75 30 30 |la.Tuple2;..\u00| > {code} > I suspect this is because the HTTP api emits the executorInfo.data directly: > {code} > JSON::Object model(const ExecutorInfo& executorInfo) > { > JSON::Object object; > object.values["executor_id"] = executorInfo.executor_id().value(); > object.values["name"] = executorInfo.name(); > object.values["data"] = executorInfo.data(); > object.values["framework_id"] = executorInfo.framework_id().value(); > object.values["command"] = model(executorInfo.command()); > object.values["resources"] = model(executorInfo.resources()); > return object; > } > {code} > I think this may be because the custom JSON processing library in stout seems > to not have any idea of what a byte array is. I'm guessing that some > implicit conversion makes it get written as a String instead, but: > {code} > inline std::ostream& operator<<(std::ostream& out, const String& string) > { > // TODO(benh): This escaping DOES NOT handle unicode, it encodes as ASCII. > // See RFC4627 for the JSON string specificiation. > return out << picojson::value(string.value).serialize(); > } > {code} > Thank you for any assistance here. Our cluster is currently entirely down -- > the frameworks cannot handle parsing the invalid JSON produced (it is not > even valid utf-8) -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (MESOS-3771) Mesos JSON API creates invalid JSON due to lack of binary data / non-ASCII handling
[ https://issues.apache.org/jira/browse/MESOS-3771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14970064#comment-14970064 ] Steven Schlansker commented on MESOS-3771: -- Sounds good to me. I think removing that field from JSON is fine for us. > Mesos JSON API creates invalid JSON due to lack of binary data / non-ASCII > handling > --- > > Key: MESOS-3771 > URL: https://issues.apache.org/jira/browse/MESOS-3771 > Project: Mesos > Issue Type: Bug > Components: HTTP API >Affects Versions: 0.24.1, 0.26.0 >Reporter: Steven Schlansker >Assignee: Joseph Wu >Priority: Critical > Labels: mesosphere > > Spark encodes some binary data into the ExecutorInfo.data field. This field > is sent as a "bytes" Protobuf value, which can have arbitrary non-UTF8 data. > If you have such a field, it seems that it is splatted out into JSON without > any regards to proper character encoding: > {code} > 0006b0b0 2e 73 70 61 72 6b 2e 65 78 65 63 75 74 6f 72 2e |.spark.executor.| > 0006b0c0 4d 65 73 6f 73 45 78 65 63 75 74 6f 72 42 61 63 |MesosExecutorBac| > 0006b0d0 6b 65 6e 64 22 7d 2c 22 64 61 74 61 22 3a 22 ac |kend"},"data":".| > 0006b0e0 ed 5c 75 30 30 30 30 5c 75 30 30 30 35 75 72 5c |.\u\u0005ur\| > 0006b0f0 75 30 30 30 30 5c 75 30 30 30 66 5b 4c 73 63 61 |u\u000f[Lsca| > 0006b100 6c 61 2e 54 75 70 6c 65 32 3b 2e cc 5c 75 30 30 |la.Tuple2;..\u00| > {code} > I suspect this is because the HTTP api emits the executorInfo.data directly: > {code} > JSON::Object model(const ExecutorInfo& executorInfo) > { > JSON::Object object; > object.values["executor_id"] = executorInfo.executor_id().value(); > object.values["name"] = executorInfo.name(); > object.values["data"] = executorInfo.data(); > object.values["framework_id"] = executorInfo.framework_id().value(); > object.values["command"] = model(executorInfo.command()); > object.values["resources"] = model(executorInfo.resources()); > return object; > } > {code} > I think this may be because the custom JSON processing library in stout seems > to not have any idea of what a byte array is. I'm guessing that some > implicit conversion makes it get written as a String instead, but: > {code} > inline std::ostream& operator<<(std::ostream& out, const String& string) > { > // TODO(benh): This escaping DOES NOT handle unicode, it encodes as ASCII. > // See RFC4627 for the JSON string specificiation. > return out << picojson::value(string.value).serialize(); > } > {code} > Thank you for any assistance here. Our cluster is currently entirely down -- > the frameworks cannot handle parsing the invalid JSON produced (it is not > even valid utf-8) -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (MESOS-3771) Mesos JSON API creates invalid JSON due to lack of binary data / non-ASCII handling
[ https://issues.apache.org/jira/browse/MESOS-3771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14970063#comment-14970063 ] Joseph Wu commented on MESOS-3771: -- Sync'd with BenH and [~bmahler] about this (offline). The proposed solution is the following: # None of the state endpoints should be dumping out the binary {{data}} fields in the first place. This includes {{ExecutorInfo}} ([dumped by master|https://github.com/apache/mesos/blob/master/src/common/http.cpp#L317]) and {{TaskInfo}} ([dumped by agent|https://github.com/apache/mesos/blob/master/src/slave/http.cpp#L103]). #* These fields should be removed from the output entirely. Existing frameworks should not be relying on this information. [~stevenschlansker], can you confirm this with Spark? #* We can easily back-port this patch, if absolutely necessary. # Master should not be storing the {{data}} fields from {{ExecutorInfo}}. We currently [store the entire object|https://github.com/apache/mesos/blob/master/src/master/master.hpp#L262-L271], which means master would be as high risk of OOM-ing if a bunch of executors were started with big {{data}} blobs. #* Master should scrub out unneeded bloat from {{ExecutorInfo}} before storing it. #* We can use an alternate internal object, like we do for {{TaskInfo}} vs {{Task}}; see [this|https://github.com/apache/mesos/blob/master/src/messages/messages.proto#L39-L41]. > Mesos JSON API creates invalid JSON due to lack of binary data / non-ASCII > handling > --- > > Key: MESOS-3771 > URL: https://issues.apache.org/jira/browse/MESOS-3771 > Project: Mesos > Issue Type: Bug > Components: HTTP API >Affects Versions: 0.24.1, 0.26.0 >Reporter: Steven Schlansker >Assignee: Joseph Wu >Priority: Critical > Labels: mesosphere > > Spark encodes some binary data into the ExecutorInfo.data field. This field > is sent as a "bytes" Protobuf value, which can have arbitrary non-UTF8 data. > If you have such a field, it seems that it is splatted out into JSON without > any regards to proper character encoding: > {code} > 0006b0b0 2e 73 70 61 72 6b 2e 65 78 65 63 75 74 6f 72 2e |.spark.executor.| > 0006b0c0 4d 65 73 6f 73 45 78 65 63 75 74 6f 72 42 61 63 |MesosExecutorBac| > 0006b0d0 6b 65 6e 64 22 7d 2c 22 64 61 74 61 22 3a 22 ac |kend"},"data":".| > 0006b0e0 ed 5c 75 30 30 30 30 5c 75 30 30 30 35 75 72 5c |.\u\u0005ur\| > 0006b0f0 75 30 30 30 30 5c 75 30 30 30 66 5b 4c 73 63 61 |u\u000f[Lsca| > 0006b100 6c 61 2e 54 75 70 6c 65 32 3b 2e cc 5c 75 30 30 |la.Tuple2;..\u00| > {code} > I suspect this is because the HTTP api emits the executorInfo.data directly: > {code} > JSON::Object model(const ExecutorInfo& executorInfo) > { > JSON::Object object; > object.values["executor_id"] = executorInfo.executor_id().value(); > object.values["name"] = executorInfo.name(); > object.values["data"] = executorInfo.data(); > object.values["framework_id"] = executorInfo.framework_id().value(); > object.values["command"] = model(executorInfo.command()); > object.values["resources"] = model(executorInfo.resources()); > return object; > } > {code} > I think this may be because the custom JSON processing library in stout seems > to not have any idea of what a byte array is. I'm guessing that some > implicit conversion makes it get written as a String instead, but: > {code} > inline std::ostream& operator<<(std::ostream& out, const String& string) > { > // TODO(benh): This escaping DOES NOT handle unicode, it encodes as ASCII. > // See RFC4627 for the JSON string specificiation. > return out << picojson::value(string.value).serialize(); > } > {code} > Thank you for any assistance here. Our cluster is currently entirely down -- > the frameworks cannot handle parsing the invalid JSON produced (it is not > even valid utf-8) -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (MESOS-3771) Mesos JSON API creates invalid JSON due to lack of binary data / non-ASCII handling
[ https://issues.apache.org/jira/browse/MESOS-3771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14968275#comment-14968275 ] Joseph Wu commented on MESOS-3771: -- Looks like our JSON library will never catch this (it's more permissive), which is why none of our unit tests have caught this. I agree that this is a problem though. I'll see if I can get more eyes on this. > Mesos JSON API creates invalid JSON due to lack of binary data / non-ASCII > handling > --- > > Key: MESOS-3771 > URL: https://issues.apache.org/jira/browse/MESOS-3771 > Project: Mesos > Issue Type: Bug > Components: HTTP API >Affects Versions: 0.24.1, 0.26.0 >Reporter: Steven Schlansker >Priority: Critical > > Spark encodes some binary data into the ExecutorInfo.data field. This field > is sent as a "bytes" Protobuf value, which can have arbitrary non-UTF8 data. > If you have such a field, it seems that it is splatted out into JSON without > any regards to proper character encoding: > {code} > 0006b0b0 2e 73 70 61 72 6b 2e 65 78 65 63 75 74 6f 72 2e |.spark.executor.| > 0006b0c0 4d 65 73 6f 73 45 78 65 63 75 74 6f 72 42 61 63 |MesosExecutorBac| > 0006b0d0 6b 65 6e 64 22 7d 2c 22 64 61 74 61 22 3a 22 ac |kend"},"data":".| > 0006b0e0 ed 5c 75 30 30 30 30 5c 75 30 30 30 35 75 72 5c |.\u\u0005ur\| > 0006b0f0 75 30 30 30 30 5c 75 30 30 30 66 5b 4c 73 63 61 |u\u000f[Lsca| > 0006b100 6c 61 2e 54 75 70 6c 65 32 3b 2e cc 5c 75 30 30 |la.Tuple2;..\u00| > {code} > I suspect this is because the HTTP api emits the executorInfo.data directly: > {code} > JSON::Object model(const ExecutorInfo& executorInfo) > { > JSON::Object object; > object.values["executor_id"] = executorInfo.executor_id().value(); > object.values["name"] = executorInfo.name(); > object.values["data"] = executorInfo.data(); > object.values["framework_id"] = executorInfo.framework_id().value(); > object.values["command"] = model(executorInfo.command()); > object.values["resources"] = model(executorInfo.resources()); > return object; > } > {code} > I think this may be because the custom JSON processing library in stout seems > to not have any idea of what a byte array is. I'm guessing that some > implicit conversion makes it get written as a String instead, but: > {code} > inline std::ostream& operator<<(std::ostream& out, const String& string) > { > // TODO(benh): This escaping DOES NOT handle unicode, it encodes as ASCII. > // See RFC4627 for the JSON string specificiation. > return out << picojson::value(string.value).serialize(); > } > {code} > Thank you for any assistance here. Our cluster is currently entirely down -- > the frameworks cannot handle parsing the invalid JSON produced (it is not > even valid utf-8) -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (MESOS-3771) Mesos JSON API creates invalid JSON due to lack of binary data / non-ASCII handling
[ https://issues.apache.org/jira/browse/MESOS-3771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14967962#comment-14967962 ] Steven Schlansker commented on MESOS-3771: -- Okay, I have distilled down the reproduction case. Using the Python test-framework with the following diff applied: {code} diff --git a/src/examples/python/test_framework.py b/src/examples/python/test_framework.py index 6af6d22..95abb97 100755 --- a/src/examples/python/test_framework.py +++ b/src/examples/python/test_framework.py @@ -150,6 +150,7 @@ class TestScheduler(mesos.interface.Scheduler): print "but received", self.messagesReceived sys.exit(1) print "All tasks done, and all messages received, exiting" +time.sleep(30) driver.stop() if __name__ == "__main__": @@ -158,6 +159,7 @@ if __name__ == "__main__": sys.exit(1) executor = mesos_pb2.ExecutorInfo() +executor.data = b'\xAC\xED' executor.executor_id.value = "default" executor.command.value = os.path.abspath("./test-executor") executor.name = "Test Executor (Python)" {code} if you run the test framework, and during the 30 second wait after it finishes, try to grab the {{/master/state.json}} endpoint, you will get a response that has invalid UTF8 in it: {code} Caused by: com.fasterxml.jackson.core.JsonParseException: Invalid UTF-8 start byte 0xac at [Source: org.jboss.netty.buffer.ChannelBufferInputStream@54c8158d; line: 1, column: 6432] {code} I tested against both 0.24.1 and current master, both exhibit the bad behavior. > Mesos JSON API creates invalid JSON due to lack of binary data / non-ASCII > handling > --- > > Key: MESOS-3771 > URL: https://issues.apache.org/jira/browse/MESOS-3771 > Project: Mesos > Issue Type: Bug > Components: HTTP API >Affects Versions: 0.24.1, 0.26.0 >Reporter: Steven Schlansker >Priority: Critical > > Spark encodes some binary data into the ExecutorInfo.data field. This field > is sent as a "bytes" Protobuf value, which can have arbitrary non-UTF8 data. > If you have such a field, it seems that it is splatted out into JSON without > any regards to proper character encoding: > {code} > 0006b0b0 2e 73 70 61 72 6b 2e 65 78 65 63 75 74 6f 72 2e |.spark.executor.| > 0006b0c0 4d 65 73 6f 73 45 78 65 63 75 74 6f 72 42 61 63 |MesosExecutorBac| > 0006b0d0 6b 65 6e 64 22 7d 2c 22 64 61 74 61 22 3a 22 ac |kend"},"data":".| > 0006b0e0 ed 5c 75 30 30 30 30 5c 75 30 30 30 35 75 72 5c |.\u\u0005ur\| > 0006b0f0 75 30 30 30 30 5c 75 30 30 30 66 5b 4c 73 63 61 |u\u000f[Lsca| > 0006b100 6c 61 2e 54 75 70 6c 65 32 3b 2e cc 5c 75 30 30 |la.Tuple2;..\u00| > {code} > I suspect this is because the HTTP api emits the executorInfo.data directly: > {code} > JSON::Object model(const ExecutorInfo& executorInfo) > { > JSON::Object object; > object.values["executor_id"] = executorInfo.executor_id().value(); > object.values["name"] = executorInfo.name(); > object.values["data"] = executorInfo.data(); > object.values["framework_id"] = executorInfo.framework_id().value(); > object.values["command"] = model(executorInfo.command()); > object.values["resources"] = model(executorInfo.resources()); > return object; > } > {code} > I think this may be because the custom JSON processing library in stout seems > to not have any idea of what a byte array is. I'm guessing that some > implicit conversion makes it get written as a String instead, but: > {code} > inline std::ostream& operator<<(std::ostream& out, const String& string) > { > // TODO(benh): This escaping DOES NOT handle unicode, it encodes as ASCII. > // See RFC4627 for the JSON string specificiation. > return out << picojson::value(string.value).serialize(); > } > {code} > Thank you for any assistance here. Our cluster is currently entirely down -- > the frameworks cannot handle parsing the invalid JSON produced (it is not > even valid utf-8) -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (MESOS-3771) Mesos JSON API creates invalid JSON due to lack of binary data / non-ASCII handling
[ https://issues.apache.org/jira/browse/MESOS-3771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14966036#comment-14966036 ] Joseph Wu commented on MESOS-3771: -- Actually, I can't repro this behavior in a unit test. ([Attempted here|https://github.com/kaysoky/mesos/commit/d8869f0aa1fdcf38072b45a6238b191c67b7e0f7]). I've constructed an {{ExecutorInfo}} with a {{data}} field holding the same data you have above. Fetching the same {{ExecutorInfo}} from the {{/state}} endpoint also gives valid JSON. Am I doing something differently? > Mesos JSON API creates invalid JSON due to lack of binary data / non-ASCII > handling > --- > > Key: MESOS-3771 > URL: https://issues.apache.org/jira/browse/MESOS-3771 > Project: Mesos > Issue Type: Bug > Components: HTTP API >Affects Versions: 0.24.1 >Reporter: Steven Schlansker >Priority: Critical > > Spark encodes some binary data into the ExecutorInfo.data field. This field > is sent as a "bytes" Protobuf value, which can have arbitrary non-UTF8 data. > If you have such a field, it seems that it is splatted out into JSON without > any regards to proper character encoding: > {code} > 0006b0b0 2e 73 70 61 72 6b 2e 65 78 65 63 75 74 6f 72 2e |.spark.executor.| > 0006b0c0 4d 65 73 6f 73 45 78 65 63 75 74 6f 72 42 61 63 |MesosExecutorBac| > 0006b0d0 6b 65 6e 64 22 7d 2c 22 64 61 74 61 22 3a 22 ac |kend"},"data":".| > 0006b0e0 ed 5c 75 30 30 30 30 5c 75 30 30 30 35 75 72 5c |.\u\u0005ur\| > 0006b0f0 75 30 30 30 30 5c 75 30 30 30 66 5b 4c 73 63 61 |u\u000f[Lsca| > 0006b100 6c 61 2e 54 75 70 6c 65 32 3b 2e cc 5c 75 30 30 |la.Tuple2;..\u00| > {code} > I suspect this is because the HTTP api emits the executorInfo.data directly: > {code} > JSON::Object model(const ExecutorInfo& executorInfo) > { > JSON::Object object; > object.values["executor_id"] = executorInfo.executor_id().value(); > object.values["name"] = executorInfo.name(); > object.values["data"] = executorInfo.data(); > object.values["framework_id"] = executorInfo.framework_id().value(); > object.values["command"] = model(executorInfo.command()); > object.values["resources"] = model(executorInfo.resources()); > return object; > } > {code} > I think this may be because the custom JSON processing library in stout seems > to not have any idea of what a byte array is. I'm guessing that some > implicit conversion makes it get written as a String instead, but: > {code} > inline std::ostream& operator<<(std::ostream& out, const String& string) > { > // TODO(benh): This escaping DOES NOT handle unicode, it encodes as ASCII. > // See RFC4627 for the JSON string specificiation. > return out << picojson::value(string.value).serialize(); > } > {code} > Thank you for any assistance here. Our cluster is currently entirely down -- > the frameworks cannot handle parsing the invalid JSON produced (it is not > even valid utf-8) -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (MESOS-3771) Mesos JSON API creates invalid JSON due to lack of binary data / non-ASCII handling
[ https://issues.apache.org/jira/browse/MESOS-3771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14965945#comment-14965945 ] Steven Schlansker commented on MESOS-3771: -- Yeah, we could try to patch Spark. However I'm sure someone else will make exactly the same mistake down the road -- it seems to work as long as you use the protobuf api only. It really seems wrong to just assume that arbitrary bytes are valid UTF-8. Note that ASCII is a real misnomer here, the only things that matter are "arbitrary binary data" (the type of 'data') and "UTF8" (the format that the rendered JSON *must* be in). I don't see anywhere here that ASCII is relevant. Maybe it's possible to escape the 0xACED sequence we see with \uXXX escapes, but I'm not sure that's possible, as those escapes produce UTF-16 codepoints not binary data... > Mesos JSON API creates invalid JSON due to lack of binary data / non-ASCII > handling > --- > > Key: MESOS-3771 > URL: https://issues.apache.org/jira/browse/MESOS-3771 > Project: Mesos > Issue Type: Bug > Components: HTTP API >Affects Versions: 0.24.1 >Reporter: Steven Schlansker >Priority: Critical > > Spark encodes some binary data into the ExecutorInfo.data field. This field > is sent as a "bytes" Protobuf value, which can have arbitrary non-UTF8 data. > If you have such a field, it seems that it is splatted out into JSON without > any regards to proper character encoding: > {code} > 0006b0b0 2e 73 70 61 72 6b 2e 65 78 65 63 75 74 6f 72 2e |.spark.executor.| > 0006b0c0 4d 65 73 6f 73 45 78 65 63 75 74 6f 72 42 61 63 |MesosExecutorBac| > 0006b0d0 6b 65 6e 64 22 7d 2c 22 64 61 74 61 22 3a 22 ac |kend"},"data":".| > 0006b0e0 ed 5c 75 30 30 30 30 5c 75 30 30 30 35 75 72 5c |.\u\u0005ur\| > 0006b0f0 75 30 30 30 30 5c 75 30 30 30 66 5b 4c 73 63 61 |u\u000f[Lsca| > 0006b100 6c 61 2e 54 75 70 6c 65 32 3b 2e cc 5c 75 30 30 |la.Tuple2;..\u00| > {code} > I suspect this is because the HTTP api emits the executorInfo.data directly: > {code} > JSON::Object model(const ExecutorInfo& executorInfo) > { > JSON::Object object; > object.values["executor_id"] = executorInfo.executor_id().value(); > object.values["name"] = executorInfo.name(); > object.values["data"] = executorInfo.data(); > object.values["framework_id"] = executorInfo.framework_id().value(); > object.values["command"] = model(executorInfo.command()); > object.values["resources"] = model(executorInfo.resources()); > return object; > } > {code} > I think this may be because the custom JSON processing library in stout seems > to not have any idea of what a byte array is. I'm guessing that some > implicit conversion makes it get written as a String instead, but: > {code} > inline std::ostream& operator<<(std::ostream& out, const String& string) > { > // TODO(benh): This escaping DOES NOT handle unicode, it encodes as ASCII. > // See RFC4627 for the JSON string specificiation. > return out << picojson::value(string.value).serialize(); > } > {code} > Thank you for any assistance here. Our cluster is currently entirely down -- > the frameworks cannot handle parsing the invalid JSON produced (it is not > even valid utf-8) -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (MESOS-3771) Mesos JSON API creates invalid JSON due to lack of binary data / non-ASCII handling
[ https://issues.apache.org/jira/browse/MESOS-3771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14965808#comment-14965808 ] Joseph Wu commented on MESOS-3771: -- ^ That's actually what would (sort of) fix your issue. There's an old TODO [here|https://github.com/apache/mesos/blob/master/src/master/http.cpp#L118-L119] to make the change. We do actually encode {{bytes}} in base64, but only when they are transformed into JSON from Protobuf. However, some of the endpoints (the ones which must be backwards compatible) appear to treat {{bytes}} as ASCII strings. If you have more control over your version of Spark, you could base64 encode from Spark: https://github.com/apache/spark/blob/master/core/src/main/scala/org/apache/spark/executor/MesosExecutorBackend.scala#L47 > Mesos JSON API creates invalid JSON due to lack of binary data / non-ASCII > handling > --- > > Key: MESOS-3771 > URL: https://issues.apache.org/jira/browse/MESOS-3771 > Project: Mesos > Issue Type: Bug > Components: HTTP API >Affects Versions: 0.24.1 >Reporter: Steven Schlansker >Priority: Critical > > Spark encodes some binary data into the ExecutorInfo.data field. This field > is sent as a "bytes" Protobuf value, which can have arbitrary non-UTF8 data. > If you have such a field, it seems that it is splatted out into JSON without > any regards to proper character encoding: > {code} > 0006b0b0 2e 73 70 61 72 6b 2e 65 78 65 63 75 74 6f 72 2e |.spark.executor.| > 0006b0c0 4d 65 73 6f 73 45 78 65 63 75 74 6f 72 42 61 63 |MesosExecutorBac| > 0006b0d0 6b 65 6e 64 22 7d 2c 22 64 61 74 61 22 3a 22 ac |kend"},"data":".| > 0006b0e0 ed 5c 75 30 30 30 30 5c 75 30 30 30 35 75 72 5c |.\u\u0005ur\| > 0006b0f0 75 30 30 30 30 5c 75 30 30 30 66 5b 4c 73 63 61 |u\u000f[Lsca| > 0006b100 6c 61 2e 54 75 70 6c 65 32 3b 2e cc 5c 75 30 30 |la.Tuple2;..\u00| > {code} > I suspect this is because the HTTP api emits the executorInfo.data directly: > {code} > JSON::Object model(const ExecutorInfo& executorInfo) > { > JSON::Object object; > object.values["executor_id"] = executorInfo.executor_id().value(); > object.values["name"] = executorInfo.name(); > object.values["data"] = executorInfo.data(); > object.values["framework_id"] = executorInfo.framework_id().value(); > object.values["command"] = model(executorInfo.command()); > object.values["resources"] = model(executorInfo.resources()); > return object; > } > {code} > I think this may be because the custom JSON processing library in stout seems > to not have any idea of what a byte array is. I'm guessing that some > implicit conversion makes it get written as a String instead, but: > {code} > inline std::ostream& operator<<(std::ostream& out, const String& string) > { > // TODO(benh): This escaping DOES NOT handle unicode, it encodes as ASCII. > // See RFC4627 for the JSON string specificiation. > return out << picojson::value(string.value).serialize(); > } > {code} > Thank you for any assistance here. Our cluster is currently entirely down -- > the frameworks cannot handle parsing the invalid JSON produced (it is not > even valid utf-8) -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (MESOS-3771) Mesos JSON API creates invalid JSON due to lack of binary data / non-ASCII handling
[ https://issues.apache.org/jira/browse/MESOS-3771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14965762#comment-14965762 ] Steven Schlansker commented on MESOS-3771: -- Similar, but potentially unrelated, issue: https://issues.apache.org/jira/browse/MESOS-3284 > Mesos JSON API creates invalid JSON due to lack of binary data / non-ASCII > handling > --- > > Key: MESOS-3771 > URL: https://issues.apache.org/jira/browse/MESOS-3771 > Project: Mesos > Issue Type: Bug > Components: HTTP API >Affects Versions: 0.24.1 >Reporter: Steven Schlansker >Priority: Critical > > Spark encodes some binary data into the ExecutorInfo.data field. This field > is sent as a "bytes" Protobuf value, which can have arbitrary non-UTF8 data. > If you have such a field, it seems that it is splatted out into JSON without > any regards to proper character encoding: > {code} > 0006b0b0 2e 73 70 61 72 6b 2e 65 78 65 63 75 74 6f 72 2e |.spark.executor.| > 0006b0c0 4d 65 73 6f 73 45 78 65 63 75 74 6f 72 42 61 63 |MesosExecutorBac| > 0006b0d0 6b 65 6e 64 22 7d 2c 22 64 61 74 61 22 3a 22 ac |kend"},"data":".| > 0006b0e0 ed 5c 75 30 30 30 30 5c 75 30 30 30 35 75 72 5c |.\u\u0005ur\| > 0006b0f0 75 30 30 30 30 5c 75 30 30 30 66 5b 4c 73 63 61 |u\u000f[Lsca| > 0006b100 6c 61 2e 54 75 70 6c 65 32 3b 2e cc 5c 75 30 30 |la.Tuple2;..\u00| > {code} > I suspect this is because the HTTP api emits the executorInfo.data directly: > {code} > JSON::Object model(const ExecutorInfo& executorInfo) > { > JSON::Object object; > object.values["executor_id"] = executorInfo.executor_id().value(); > object.values["name"] = executorInfo.name(); > object.values["data"] = executorInfo.data(); > object.values["framework_id"] = executorInfo.framework_id().value(); > object.values["command"] = model(executorInfo.command()); > object.values["resources"] = model(executorInfo.resources()); > return object; > } > {code} > I think this may be because the custom JSON processing library in stout seems > to not have any idea of what a byte array is. I'm guessing that some > implicit conversion makes it get written as a String instead, but: > {code} > inline std::ostream& operator<<(std::ostream& out, const String& string) > { > // TODO(benh): This escaping DOES NOT handle unicode, it encodes as ASCII. > // See RFC4627 for the JSON string specificiation. > return out << picojson::value(string.value).serialize(); > } > {code} > Thank you for any assistance here. Our cluster is currently entirely down -- > the frameworks cannot handle parsing the invalid JSON produced (it is not > even valid utf-8) -- This message was sent by Atlassian JIRA (v6.3.4#6332)