Re: When does Spark switch from PROCESS_LOCAL to NODE_LOCAL or RACK_LOCAL?
Hi Brad and Nick, Thanks for the comments! I opened a ticket to get a more thorough explanation of data locality into the docs here: https://issues.apache.org/jira/browse/SPARK-3526 If you could put any other unanswered questions you have about data locality on that ticket I'll try to incorporate answers to them in the final addition I send in. Andrew On Sun, Sep 14, 2014 at 6:47 PM, Brad Miller wrote: > Hi Andrew, > > I agree with Nicholas. That was a nice, concise summary of the > meaning of the locality customization options, indicators and default > Spark behaviors. I haven't combed through the documentation > end-to-end in a while, but I'm also not sure that information is > presently represented somewhere and it would be great to persist it > somewhere besides the mailing list. > > best, > -Brad > > On Fri, Sep 12, 2014 at 12:12 PM, Nicholas Chammas > wrote: > > Andrew, > > > > This email was pretty helpful. I feel like this stuff should be > summarized > > in the docs somewhere, or perhaps in a blog post. > > > > Do you know if it is? > > > > Nick > > > > > > On Thu, Jun 5, 2014 at 6:36 PM, Andrew Ash wrote: > >> > >> The locality is how close the data is to the code that's processing it. > >> PROCESS_LOCAL means data is in the same JVM as the code that's running, > so > >> it's really fast. NODE_LOCAL might mean that the data is in HDFS on the > >> same node, or in another executor on the same node, so is a little > slower > >> because the data has to travel across an IPC connection. RACK_LOCAL is > even > >> slower -- data is on a different server so needs to be sent over the > >> network. > >> > >> Spark switches to lower locality levels when there's no unprocessed data > >> on a node that has idle CPUs. In that situation you have two options: > wait > >> until the busy CPUs free up so you can start another task that uses > data on > >> that server, or start a new task on a farther away server that needs to > >> bring data from that remote place. What Spark typically does is wait a > bit > >> in the hopes that a busy CPU frees up. Once that timeout expires, it > starts > >> moving the data from far away to the free CPU. > >> > >> The main tunable option is how far long the scheduler waits before > >> starting to move data rather than code. Those are the spark.locality.* > >> settings here: http://spark.apache.org/docs/latest/configuration.html > >> > >> If you want to prevent this from happening entirely, you can set the > >> values to ridiculously high numbers. The documentation also mentions > that > >> "0" has special meaning, so you can try that as well. > >> > >> Good luck! > >> Andrew > >> > >> > >> On Thu, Jun 5, 2014 at 3:13 PM, Sung Hwan Chung < > coded...@cs.stanford.edu> > >> wrote: > >>> > >>> I noticed that sometimes tasks would switch from PROCESS_LOCAL (I'd > >>> assume that this means fully cached) to NODE_LOCAL or even RACK_LOCAL. > >>> > >>> When these happen things get extremely slow. > >>> > >>> Does this mean that the executor got terminated and restarted? > >>> > >>> Is there a way to prevent this from happening (barring the machine > >>> actually going down, I'd rather stick with the same process)? > >> > >> > > > > - > To unsubscribe, e-mail: user-unsubscr...@spark.apache.org > For additional commands, e-mail: user-h...@spark.apache.org > >
Re: When does Spark switch from PROCESS_LOCAL to NODE_LOCAL or RACK_LOCAL?
Hi Andrew, I agree with Nicholas. That was a nice, concise summary of the meaning of the locality customization options, indicators and default Spark behaviors. I haven't combed through the documentation end-to-end in a while, but I'm also not sure that information is presently represented somewhere and it would be great to persist it somewhere besides the mailing list. best, -Brad On Fri, Sep 12, 2014 at 12:12 PM, Nicholas Chammas wrote: > Andrew, > > This email was pretty helpful. I feel like this stuff should be summarized > in the docs somewhere, or perhaps in a blog post. > > Do you know if it is? > > Nick > > > On Thu, Jun 5, 2014 at 6:36 PM, Andrew Ash wrote: >> >> The locality is how close the data is to the code that's processing it. >> PROCESS_LOCAL means data is in the same JVM as the code that's running, so >> it's really fast. NODE_LOCAL might mean that the data is in HDFS on the >> same node, or in another executor on the same node, so is a little slower >> because the data has to travel across an IPC connection. RACK_LOCAL is even >> slower -- data is on a different server so needs to be sent over the >> network. >> >> Spark switches to lower locality levels when there's no unprocessed data >> on a node that has idle CPUs. In that situation you have two options: wait >> until the busy CPUs free up so you can start another task that uses data on >> that server, or start a new task on a farther away server that needs to >> bring data from that remote place. What Spark typically does is wait a bit >> in the hopes that a busy CPU frees up. Once that timeout expires, it starts >> moving the data from far away to the free CPU. >> >> The main tunable option is how far long the scheduler waits before >> starting to move data rather than code. Those are the spark.locality.* >> settings here: http://spark.apache.org/docs/latest/configuration.html >> >> If you want to prevent this from happening entirely, you can set the >> values to ridiculously high numbers. The documentation also mentions that >> "0" has special meaning, so you can try that as well. >> >> Good luck! >> Andrew >> >> >> On Thu, Jun 5, 2014 at 3:13 PM, Sung Hwan Chung >> wrote: >>> >>> I noticed that sometimes tasks would switch from PROCESS_LOCAL (I'd >>> assume that this means fully cached) to NODE_LOCAL or even RACK_LOCAL. >>> >>> When these happen things get extremely slow. >>> >>> Does this mean that the executor got terminated and restarted? >>> >>> Is there a way to prevent this from happening (barring the machine >>> actually going down, I'd rather stick with the same process)? >> >> > - To unsubscribe, e-mail: user-unsubscr...@spark.apache.org For additional commands, e-mail: user-h...@spark.apache.org
Re: When does Spark switch from PROCESS_LOCAL to NODE_LOCAL or RACK_LOCAL?
Another observation I had was reading over local filesystem with “file://“. it was stated as PROCESS_LOCAL which was confusing. Regards, Liming On 13 Sep, 2014, at 3:12 am, Nicholas Chammas wrote: > Andrew, > > This email was pretty helpful. I feel like this stuff should be summarized in > the docs somewhere, or perhaps in a blog post. > > Do you know if it is? > > Nick > > > On Thu, Jun 5, 2014 at 6:36 PM, Andrew Ash wrote: > The locality is how close the data is to the code that's processing it. > PROCESS_LOCAL means data is in the same JVM as the code that's running, so > it's really fast. NODE_LOCAL might mean that the data is in HDFS on the same > node, or in another executor on the same node, so is a little slower because > the data has to travel across an IPC connection. RACK_LOCAL is even slower > -- data is on a different server so needs to be sent over the network. > > Spark switches to lower locality levels when there's no unprocessed data on a > node that has idle CPUs. In that situation you have two options: wait until > the busy CPUs free up so you can start another task that uses data on that > server, or start a new task on a farther away server that needs to bring data > from that remote place. What Spark typically does is wait a bit in the hopes > that a busy CPU frees up. Once that timeout expires, it starts moving the > data from far away to the free CPU. > > The main tunable option is how far long the scheduler waits before starting > to move data rather than code. Those are the spark.locality.* settings here: > http://spark.apache.org/docs/latest/configuration.html > > If you want to prevent this from happening entirely, you can set the values > to ridiculously high numbers. The documentation also mentions that "0" has > special meaning, so you can try that as well. > > Good luck! > Andrew > > > On Thu, Jun 5, 2014 at 3:13 PM, Sung Hwan Chung > wrote: > I noticed that sometimes tasks would switch from PROCESS_LOCAL (I'd assume > that this means fully cached) to NODE_LOCAL or even RACK_LOCAL. > > When these happen things get extremely slow. > > Does this mean that the executor got terminated and restarted? > > Is there a way to prevent this from happening (barring the machine actually > going down, I'd rather stick with the same process)? > >
Re: When does Spark switch from PROCESS_LOCAL to NODE_LOCAL or RACK_LOCAL?
Andrew, This email was pretty helpful. I feel like this stuff should be summarized in the docs somewhere, or perhaps in a blog post. Do you know if it is? Nick On Thu, Jun 5, 2014 at 6:36 PM, Andrew Ash wrote: > The locality is how close the data is to the code that's processing it. > PROCESS_LOCAL means data is in the same JVM as the code that's running, so > it's really fast. NODE_LOCAL might mean that the data is in HDFS on the > same node, or in another executor on the same node, so is a little slower > because the data has to travel across an IPC connection. RACK_LOCAL is > even slower -- data is on a different server so needs to be sent over the > network. > > Spark switches to lower locality levels when there's no unprocessed data > on a node that has idle CPUs. In that situation you have two options: wait > until the busy CPUs free up so you can start another task that uses data on > that server, or start a new task on a farther away server that needs to > bring data from that remote place. What Spark typically does is wait a bit > in the hopes that a busy CPU frees up. Once that timeout expires, it > starts moving the data from far away to the free CPU. > > The main tunable option is how far long the scheduler waits before > starting to move data rather than code. Those are the spark.locality.* > settings here: http://spark.apache.org/docs/latest/configuration.html > > If you want to prevent this from happening entirely, you can set the > values to ridiculously high numbers. The documentation also mentions that > "0" has special meaning, so you can try that as well. > > Good luck! > Andrew > > > On Thu, Jun 5, 2014 at 3:13 PM, Sung Hwan Chung > wrote: > >> I noticed that sometimes tasks would switch from PROCESS_LOCAL (I'd >> assume that this means fully cached) to NODE_LOCAL or even RACK_LOCAL. >> >> When these happen things get extremely slow. >> >> Does this mean that the executor got terminated and restarted? >> >> Is there a way to prevent this from happening (barring the machine >> actually going down, I'd rather stick with the same process)? >> > >
RE: When does Spark switch from PROCESS_LOCAL to NODE_LOCAL or RACK_LOCAL?
If some task have no locality preference, it will also show up as PROCESS_LOCAL, yet, I think we probably need to name it NO_PREFER to make it more clear. Not sure is this your case. Best Regards, Raymond Liu From: coded...@gmail.com [mailto:coded...@gmail.com] On Behalf Of Sung Hwan Chung Sent: Friday, June 06, 2014 6:53 AM To: user@spark.apache.org Subject: Re: When does Spark switch from PROCESS_LOCAL to NODE_LOCAL or RACK_LOCAL? Additionally, I've encountered some confusing situation where the locality level for a task showed up as 'PROCESS_LOCAL' even though I didn't cache the data. I wonder some implicit caching happens even without the user specifying things. On Thu, Jun 5, 2014 at 3:50 PM, Sung Hwan Chung mailto:coded...@cs.stanford.edu>> wrote: Thanks Andrew, Is there a chance that even with full-caching, that modes other than PROCESS_LOCAL will be used? E.g., let's say, an executor will try to perform tasks although the data are cached on a different executor. What I'd like to do is to prevent such a scenario entirely. I'd like to know if setting 'spark.locality.wait' to a very high value would guarantee that the mode will always be 'PROCESS_LOCAL'. On Thu, Jun 5, 2014 at 3:36 PM, Andrew Ash mailto:and...@andrewash.com>> wrote: The locality is how close the data is to the code that's processing it. PROCESS_LOCAL means data is in the same JVM as the code that's running, so it's really fast. NODE_LOCAL might mean that the data is in HDFS on the same node, or in another executor on the same node, so is a little slower because the data has to travel across an IPC connection. RACK_LOCAL is even slower -- data is on a different server so needs to be sent over the network. Spark switches to lower locality levels when there's no unprocessed data on a node that has idle CPUs. In that situation you have two options: wait until the busy CPUs free up so you can start another task that uses data on that server, or start a new task on a farther away server that needs to bring data from that remote place. What Spark typically does is wait a bit in the hopes that a busy CPU frees up. Once that timeout expires, it starts moving the data from far away to the free CPU. The main tunable option is how far long the scheduler waits before starting to move data rather than code. Those are the spark.locality.* settings here: http://spark.apache.org/docs/latest/configuration.html If you want to prevent this from happening entirely, you can set the values to ridiculously high numbers. The documentation also mentions that "0" has special meaning, so you can try that as well. Good luck! Andrew On Thu, Jun 5, 2014 at 3:13 PM, Sung Hwan Chung mailto:coded...@cs.stanford.edu>> wrote: I noticed that sometimes tasks would switch from PROCESS_LOCAL (I'd assume that this means fully cached) to NODE_LOCAL or even RACK_LOCAL. When these happen things get extremely slow. Does this mean that the executor got terminated and restarted? Is there a way to prevent this from happening (barring the machine actually going down, I'd rather stick with the same process)?
Re: When does Spark switch from PROCESS_LOCAL to NODE_LOCAL or RACK_LOCAL?
Additionally, I've encountered some confusing situation where the locality level for a task showed up as 'PROCESS_LOCAL' even though I didn't cache the data. I wonder some implicit caching happens even without the user specifying things. On Thu, Jun 5, 2014 at 3:50 PM, Sung Hwan Chung wrote: > Thanks Andrew, > > Is there a chance that even with full-caching, that modes other than > PROCESS_LOCAL will be used? E.g., let's say, an executor will try to > perform tasks although the data are cached on a different executor. > > What I'd like to do is to prevent such a scenario entirely. > > I'd like to know if setting 'spark.locality.wait' to a very high value > would guarantee that the mode will always be 'PROCESS_LOCAL'. > > > On Thu, Jun 5, 2014 at 3:36 PM, Andrew Ash wrote: > >> The locality is how close the data is to the code that's processing it. >> PROCESS_LOCAL means data is in the same JVM as the code that's running, so >> it's really fast. NODE_LOCAL might mean that the data is in HDFS on the >> same node, or in another executor on the same node, so is a little slower >> because the data has to travel across an IPC connection. RACK_LOCAL is >> even slower -- data is on a different server so needs to be sent over the >> network. >> >> Spark switches to lower locality levels when there's no unprocessed data >> on a node that has idle CPUs. In that situation you have two options: wait >> until the busy CPUs free up so you can start another task that uses data on >> that server, or start a new task on a farther away server that needs to >> bring data from that remote place. What Spark typically does is wait a bit >> in the hopes that a busy CPU frees up. Once that timeout expires, it >> starts moving the data from far away to the free CPU. >> >> The main tunable option is how far long the scheduler waits before >> starting to move data rather than code. Those are the spark.locality.* >> settings here: http://spark.apache.org/docs/latest/configuration.html >> >> If you want to prevent this from happening entirely, you can set the >> values to ridiculously high numbers. The documentation also mentions that >> "0" has special meaning, so you can try that as well. >> >> Good luck! >> Andrew >> >> >> On Thu, Jun 5, 2014 at 3:13 PM, Sung Hwan Chung > > wrote: >> >>> I noticed that sometimes tasks would switch from PROCESS_LOCAL (I'd >>> assume that this means fully cached) to NODE_LOCAL or even RACK_LOCAL. >>> >>> When these happen things get extremely slow. >>> >>> Does this mean that the executor got terminated and restarted? >>> >>> Is there a way to prevent this from happening (barring the machine >>> actually going down, I'd rather stick with the same process)? >>> >> >> >
Re: When does Spark switch from PROCESS_LOCAL to NODE_LOCAL or RACK_LOCAL?
Thanks Andrew, Is there a chance that even with full-caching, that modes other than PROCESS_LOCAL will be used? E.g., let's say, an executor will try to perform tasks although the data are cached on a different executor. What I'd like to do is to prevent such a scenario entirely. I'd like to know if setting 'spark.locality.wait' to a very high value would guarantee that the mode will always be 'PROCESS_LOCAL'. On Thu, Jun 5, 2014 at 3:36 PM, Andrew Ash wrote: > The locality is how close the data is to the code that's processing it. > PROCESS_LOCAL means data is in the same JVM as the code that's running, so > it's really fast. NODE_LOCAL might mean that the data is in HDFS on the > same node, or in another executor on the same node, so is a little slower > because the data has to travel across an IPC connection. RACK_LOCAL is > even slower -- data is on a different server so needs to be sent over the > network. > > Spark switches to lower locality levels when there's no unprocessed data > on a node that has idle CPUs. In that situation you have two options: wait > until the busy CPUs free up so you can start another task that uses data on > that server, or start a new task on a farther away server that needs to > bring data from that remote place. What Spark typically does is wait a bit > in the hopes that a busy CPU frees up. Once that timeout expires, it > starts moving the data from far away to the free CPU. > > The main tunable option is how far long the scheduler waits before > starting to move data rather than code. Those are the spark.locality.* > settings here: http://spark.apache.org/docs/latest/configuration.html > > If you want to prevent this from happening entirely, you can set the > values to ridiculously high numbers. The documentation also mentions that > "0" has special meaning, so you can try that as well. > > Good luck! > Andrew > > > On Thu, Jun 5, 2014 at 3:13 PM, Sung Hwan Chung > wrote: > >> I noticed that sometimes tasks would switch from PROCESS_LOCAL (I'd >> assume that this means fully cached) to NODE_LOCAL or even RACK_LOCAL. >> >> When these happen things get extremely slow. >> >> Does this mean that the executor got terminated and restarted? >> >> Is there a way to prevent this from happening (barring the machine >> actually going down, I'd rather stick with the same process)? >> > >
Re: When does Spark switch from PROCESS_LOCAL to NODE_LOCAL or RACK_LOCAL?
The locality is how close the data is to the code that's processing it. PROCESS_LOCAL means data is in the same JVM as the code that's running, so it's really fast. NODE_LOCAL might mean that the data is in HDFS on the same node, or in another executor on the same node, so is a little slower because the data has to travel across an IPC connection. RACK_LOCAL is even slower -- data is on a different server so needs to be sent over the network. Spark switches to lower locality levels when there's no unprocessed data on a node that has idle CPUs. In that situation you have two options: wait until the busy CPUs free up so you can start another task that uses data on that server, or start a new task on a farther away server that needs to bring data from that remote place. What Spark typically does is wait a bit in the hopes that a busy CPU frees up. Once that timeout expires, it starts moving the data from far away to the free CPU. The main tunable option is how far long the scheduler waits before starting to move data rather than code. Those are the spark.locality.* settings here: http://spark.apache.org/docs/latest/configuration.html If you want to prevent this from happening entirely, you can set the values to ridiculously high numbers. The documentation also mentions that "0" has special meaning, so you can try that as well. Good luck! Andrew On Thu, Jun 5, 2014 at 3:13 PM, Sung Hwan Chung wrote: > I noticed that sometimes tasks would switch from PROCESS_LOCAL (I'd assume > that this means fully cached) to NODE_LOCAL or even RACK_LOCAL. > > When these happen things get extremely slow. > > Does this mean that the executor got terminated and restarted? > > Is there a way to prevent this from happening (barring the machine > actually going down, I'd rather stick with the same process)? >
Re: When does Spark switch from PROCESS_LOCAL to NODE_LOCAL or RACK_LOCAL?
On a related note, I'd also minimize any kind of executor movement. I.e., once an executor is spawned and data cached in the executor, I want that executor to live all the way till the job is finished, or the machine fails in a fatal manner. What would be the best way to ensure that this is the case? On Thu, Jun 5, 2014 at 3:13 PM, Sung Hwan Chung wrote: > I noticed that sometimes tasks would switch from PROCESS_LOCAL (I'd assume > that this means fully cached) to NODE_LOCAL or even RACK_LOCAL. > > When these happen things get extremely slow. > > Does this mean that the executor got terminated and restarted? > > Is there a way to prevent this from happening (barring the machine > actually going down, I'd rather stick with the same process)? >
