You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-user@hadoop.apache.org by "O'Neill, Paul" <On...@dnb.com.INVALID> on 2016/01/08 20:44:29 UTC
Run Yarn/Hadoop jar without first unjarring to /tmp folder?
Hey,
Is it possible to execute a jar without unjarring ro /tmp first?
We have large Shaded jars are so are running out of inode space on /tmp.
I can execute without unpacking via java -jar and adding the output from "yarn classpath" however I am missing all the options that are set via the YARN script.
Is there anyway to do this other than manually duplicating the entries in yarn script?
Also, is it possible to change where the jars are unpacked to without also changing the Hadoop.tmp.dir across all the data nodes?
It seems like these should be unrelated properties?
Cheers.
[cid:image001.png@01D05044.5C2AEE60]
Paul O'Neill
Principal Engineer
The Chase Building, 5th Floor
Carmanhall Road, Sandyford,
Dublin 18, Ireland
Direct +353 1 256 6090
Mobile +353 86 835 4084
http://www.dnb.co.uk/
[cid:image002.png@01D05044.5C2AEE60]<http://www.facebook.com/DunBradstreet>[cid:image003.png@01D05044.5C2AEE60]<http://twitter.com/dnbus>[cid:image004.png@01D05044.5C2AEE60]<http://www.linkedin.com/company/dun-&-bradstreet>[cid:image005.png@01D05044.5C2AEE60]<http://www.youtube.com/user/DunandBrad>
[cid:image014.png@01D05B13.3C473430]
The information contained in this electronic message and any attachments (the "Message") is intended for one or more specific individuals or entities, and may be confidential, proprietary, privileged or otherwise protected by law. If you are not the intended recipient (or you are not authorised to receive for the recipient), please notify the sender immediately, delete this Message and do not disclose, distribute, or copy it to any third party or otherwise use this Message. Electronic messages are not secure or error free and can contain viruses or may be delayed and the sender is not liable for any of these occurrences. The sender reserves the right to monitor, record, transfer cross border and retain electronic messages.
"D&B" is a trading style of D&B Business Information Solutions is registered in Ireland. www.dnb.co.uk
Re: Run Yarn/Hadoop jar without first unjarring to /tmp folder?
Posted by "O'Neill, Paul" <On...@dnb.com.INVALID>.
Thanks Annabel,
That link seems to be specific to spark, how do I do this with yarn?
From: Annabel Melongo
Sent: Friday 8 January 21:58
Subject: Re: Run Yarn/Hadoop jar without first unjarring to /tmp folder?
To: user@hadoop.apache.org
O'Neill,
Yes, you can submit your job with jars without unpacking and depending on your cluster environment, the jars will or will not be automatically copied. This topic was discussed in this mailing list about two weeks ago. Please read Andrew Or's explanations, on Tuesday Dec. 20th, 2015, on how the Spark framework accomplishes this:
Re: Can't submit job to stand alone cluster<https://www.mail-archive.com/user@spark.apache.org/msg43223.html>
Re: Can't submit job to stand alone cluster<https://www.mail-archive.com/user@spark.apache.org/msg43223.html>
Let me clarify a few things for everyone: There are three *cluster managers*: standalone, YARN, and Mesos. Each cluster manager can run in two *deploy modes*, clie...
View on www.mail-archive.com<https://www.mail-archive.com/user@spark.apache.org/msg43223.html>
Preview by Yahoo
On Friday, January 8, 2016 2:44 PM, "O'Neill, Paul" <On...@dnb.com.INVALID> wrote:
Hey,
Is it possible to execute a jar without unjarring ro /tmp first?
We have large Shaded jars are so are running out of inode space on /tmp.
I can execute without unpacking via java -jar and adding the output from "yarn classpath" however I am missing all the options that are set via the YARN script.
Is there anyway to do this other than manually duplicating the entries in yarn script?
Also, is it possible to change where the jars are unpacked to without also changing the Hadoop.tmp.dir across all the data nodes?
It seems like these should be unrelated properties?
Cheers.
Paul O'Neill
Principal Engineer
The Chase Building, 5th Floor
Carmanhall Road, Sandyford,
Dublin 18, Ireland
Direct +353 1 256 6090
Mobile +353 86 835 4084
http://www.dnb.co.uk/
The information contained in this electronic message and any attachments (the "Message") is intended for one or more specific individuals or entities, and may be confidential, proprietary, privileged or otherwise protected by law. If you are not the intended recipient (or you are not authorised to receive for the recipient), please notify the sender immediately, delete this Message and do not disclose, distribute, or copy it to any third party or otherwise use this Message. Electronic messages are not secure or error free and can contain viruses or may be delayed and the sender is not liable for any of these occurrences. The sender reserves the right to monitor, record, transfer cross border and retain electronic messages.
"D&B" is a trading style of D&B Business Information Solutions is registered in Ireland. www.dnb.co.uk
Re: Run Yarn/Hadoop jar without first unjarring to /tmp folder?
Posted by "O'Neill, Paul" <On...@dnb.com.INVALID>.
Thanks Annabel,
That link seems to be specific to spark, how do I do this with yarn?
From: Annabel Melongo
Sent: Friday 8 January 21:58
Subject: Re: Run Yarn/Hadoop jar without first unjarring to /tmp folder?
To: user@hadoop.apache.org
O'Neill,
Yes, you can submit your job with jars without unpacking and depending on your cluster environment, the jars will or will not be automatically copied. This topic was discussed in this mailing list about two weeks ago. Please read Andrew Or's explanations, on Tuesday Dec. 20th, 2015, on how the Spark framework accomplishes this:
Re: Can't submit job to stand alone cluster<https://www.mail-archive.com/user@spark.apache.org/msg43223.html>
Re: Can't submit job to stand alone cluster<https://www.mail-archive.com/user@spark.apache.org/msg43223.html>
Let me clarify a few things for everyone: There are three *cluster managers*: standalone, YARN, and Mesos. Each cluster manager can run in two *deploy modes*, clie...
View on www.mail-archive.com<https://www.mail-archive.com/user@spark.apache.org/msg43223.html>
Preview by Yahoo
On Friday, January 8, 2016 2:44 PM, "O'Neill, Paul" <On...@dnb.com.INVALID> wrote:
Hey,
Is it possible to execute a jar without unjarring ro /tmp first?
We have large Shaded jars are so are running out of inode space on /tmp.
I can execute without unpacking via java -jar and adding the output from "yarn classpath" however I am missing all the options that are set via the YARN script.
Is there anyway to do this other than manually duplicating the entries in yarn script?
Also, is it possible to change where the jars are unpacked to without also changing the Hadoop.tmp.dir across all the data nodes?
It seems like these should be unrelated properties?
Cheers.
Paul O'Neill
Principal Engineer
The Chase Building, 5th Floor
Carmanhall Road, Sandyford,
Dublin 18, Ireland
Direct +353 1 256 6090
Mobile +353 86 835 4084
http://www.dnb.co.uk/
The information contained in this electronic message and any attachments (the "Message") is intended for one or more specific individuals or entities, and may be confidential, proprietary, privileged or otherwise protected by law. If you are not the intended recipient (or you are not authorised to receive for the recipient), please notify the sender immediately, delete this Message and do not disclose, distribute, or copy it to any third party or otherwise use this Message. Electronic messages are not secure or error free and can contain viruses or may be delayed and the sender is not liable for any of these occurrences. The sender reserves the right to monitor, record, transfer cross border and retain electronic messages.
"D&B" is a trading style of D&B Business Information Solutions is registered in Ireland. www.dnb.co.uk
Re: Run Yarn/Hadoop jar without first unjarring to /tmp folder?
Posted by "O'Neill, Paul" <On...@dnb.com.INVALID>.
Thanks Annabel,
That link seems to be specific to spark, how do I do this with yarn?
From: Annabel Melongo
Sent: Friday 8 January 21:58
Subject: Re: Run Yarn/Hadoop jar without first unjarring to /tmp folder?
To: user@hadoop.apache.org
O'Neill,
Yes, you can submit your job with jars without unpacking and depending on your cluster environment, the jars will or will not be automatically copied. This topic was discussed in this mailing list about two weeks ago. Please read Andrew Or's explanations, on Tuesday Dec. 20th, 2015, on how the Spark framework accomplishes this:
Re: Can't submit job to stand alone cluster<https://www.mail-archive.com/user@spark.apache.org/msg43223.html>
Re: Can't submit job to stand alone cluster<https://www.mail-archive.com/user@spark.apache.org/msg43223.html>
Let me clarify a few things for everyone: There are three *cluster managers*: standalone, YARN, and Mesos. Each cluster manager can run in two *deploy modes*, clie...
View on www.mail-archive.com<https://www.mail-archive.com/user@spark.apache.org/msg43223.html>
Preview by Yahoo
On Friday, January 8, 2016 2:44 PM, "O'Neill, Paul" <On...@dnb.com.INVALID> wrote:
Hey,
Is it possible to execute a jar without unjarring ro /tmp first?
We have large Shaded jars are so are running out of inode space on /tmp.
I can execute without unpacking via java -jar and adding the output from "yarn classpath" however I am missing all the options that are set via the YARN script.
Is there anyway to do this other than manually duplicating the entries in yarn script?
Also, is it possible to change where the jars are unpacked to without also changing the Hadoop.tmp.dir across all the data nodes?
It seems like these should be unrelated properties?
Cheers.
Paul O'Neill
Principal Engineer
The Chase Building, 5th Floor
Carmanhall Road, Sandyford,
Dublin 18, Ireland
Direct +353 1 256 6090
Mobile +353 86 835 4084
http://www.dnb.co.uk/
The information contained in this electronic message and any attachments (the "Message") is intended for one or more specific individuals or entities, and may be confidential, proprietary, privileged or otherwise protected by law. If you are not the intended recipient (or you are not authorised to receive for the recipient), please notify the sender immediately, delete this Message and do not disclose, distribute, or copy it to any third party or otherwise use this Message. Electronic messages are not secure or error free and can contain viruses or may be delayed and the sender is not liable for any of these occurrences. The sender reserves the right to monitor, record, transfer cross border and retain electronic messages.
"D&B" is a trading style of D&B Business Information Solutions is registered in Ireland. www.dnb.co.uk
Re: Run Yarn/Hadoop jar without first unjarring to /tmp folder?
Posted by "O'Neill, Paul" <On...@dnb.com.INVALID>.
Thanks Annabel,
That link seems to be specific to spark, how do I do this with yarn?
From: Annabel Melongo
Sent: Friday 8 January 21:58
Subject: Re: Run Yarn/Hadoop jar without first unjarring to /tmp folder?
To: user@hadoop.apache.org
O'Neill,
Yes, you can submit your job with jars without unpacking and depending on your cluster environment, the jars will or will not be automatically copied. This topic was discussed in this mailing list about two weeks ago. Please read Andrew Or's explanations, on Tuesday Dec. 20th, 2015, on how the Spark framework accomplishes this:
Re: Can't submit job to stand alone cluster<https://www.mail-archive.com/user@spark.apache.org/msg43223.html>
Re: Can't submit job to stand alone cluster<https://www.mail-archive.com/user@spark.apache.org/msg43223.html>
Let me clarify a few things for everyone: There are three *cluster managers*: standalone, YARN, and Mesos. Each cluster manager can run in two *deploy modes*, clie...
View on www.mail-archive.com<https://www.mail-archive.com/user@spark.apache.org/msg43223.html>
Preview by Yahoo
On Friday, January 8, 2016 2:44 PM, "O'Neill, Paul" <On...@dnb.com.INVALID> wrote:
Hey,
Is it possible to execute a jar without unjarring ro /tmp first?
We have large Shaded jars are so are running out of inode space on /tmp.
I can execute without unpacking via java -jar and adding the output from "yarn classpath" however I am missing all the options that are set via the YARN script.
Is there anyway to do this other than manually duplicating the entries in yarn script?
Also, is it possible to change where the jars are unpacked to without also changing the Hadoop.tmp.dir across all the data nodes?
It seems like these should be unrelated properties?
Cheers.
Paul O'Neill
Principal Engineer
The Chase Building, 5th Floor
Carmanhall Road, Sandyford,
Dublin 18, Ireland
Direct +353 1 256 6090
Mobile +353 86 835 4084
http://www.dnb.co.uk/
The information contained in this electronic message and any attachments (the "Message") is intended for one or more specific individuals or entities, and may be confidential, proprietary, privileged or otherwise protected by law. If you are not the intended recipient (or you are not authorised to receive for the recipient), please notify the sender immediately, delete this Message and do not disclose, distribute, or copy it to any third party or otherwise use this Message. Electronic messages are not secure or error free and can contain viruses or may be delayed and the sender is not liable for any of these occurrences. The sender reserves the right to monitor, record, transfer cross border and retain electronic messages.
"D&B" is a trading style of D&B Business Information Solutions is registered in Ireland. www.dnb.co.uk
Re: Run Yarn/Hadoop jar without first unjarring to /tmp folder?
Posted by Annabel Melongo <me...@yahoo.com.INVALID>.
O'Neill,
Yes, you can submit your job with jars without unpacking and depending on your cluster environment, the jars will or will not be automatically copied. This topic was discussed in this mailing list about two weeks ago. Please read Andrew Or's explanations, on Tuesday Dec. 20th, 2015, on how the Spark framework accomplishes this:
Re: Can't submit job to stand alone cluster
| |
| | | | | | | |
| Re: Can't submit job to stand alone clusterLet me clarify a few things for everyone: There are three *cluster managers*: standalone, YARN, and Mesos. Each cluster manager can run in two *deploy modes*, clie... |
| |
| View on www.mail-archive.com | Preview by Yahoo |
| |
|
|
On Friday, January 8, 2016 2:44 PM, "O'Neill, Paul" <On...@dnb.com.INVALID> wrote:
<!--#yiv5999690184 _filtered #yiv5999690184 {font-family:"Cambria Math";panose-1:2 4 5 3 5 4 6 3 2 4;} _filtered #yiv5999690184 {font-family:Calibri;panose-1:2 15 5 2 2 2 4 3 2 4;} _filtered #yiv5999690184 {font-family:"Gill Sans MT";panose-1:2 11 5 2 2 1 4 2 2 3;}#yiv5999690184 #yiv5999690184 p.yiv5999690184MsoNormal, #yiv5999690184 li.yiv5999690184MsoNormal, #yiv5999690184 div.yiv5999690184MsoNormal {margin:0cm;margin-bottom:.0001pt;font-size:11.0pt;font-family:"Calibri", sans-serif;}#yiv5999690184 a:link, #yiv5999690184 span.yiv5999690184MsoHyperlink {color:#0563C1;text-decoration:underline;}#yiv5999690184 a:visited, #yiv5999690184 span.yiv5999690184MsoHyperlinkFollowed {color:#954F72;text-decoration:underline;}#yiv5999690184 span.yiv5999690184EmailStyle17 {font-family:"Calibri", sans-serif;color:windowtext;}#yiv5999690184 .yiv5999690184MsoChpDefault {font-family:"Calibri", sans-serif;} _filtered #yiv5999690184 {margin:72.0pt 72.0pt 72.0pt 72.0pt;}#yiv5999690184 div.yiv5999690184WordSection1 {}-->Hey, Is it possible to execute a jar without unjarring ro /tmp first? We have large Shaded jars are so are running out of inode space on /tmp. I can execute without unpacking via java –jar and adding the output from “yarn classpath” however I am missing all the options that are set via the YARN script. Is there anyway to do this other than manually duplicating the entries in yarn script? Also, is it possible to change where the jars are unpacked to without also changing the Hadoop.tmp.dir across all the data nodes? It seems like these should be unrelated properties? Cheers. Paul O’Neill Principal Engineer The Chase Building, 5th Floor Carmanhall Road, Sandyford, Dublin 18, Ireland Direct +353 1 256 6090 Mobile +353 86 835 4084 http://www.dnb.co.uk/ The information contained in this electronic message and any attachments (the "Message") is intended for one or more specific individuals or entities, and may be confidential, proprietary, privileged or otherwise protected by law. If you are not the intended recipient (or you are not authorised to receive for the recipient), please notify the sender immediately, delete this Message and do not disclose, distribute, or copy it to any third party or otherwise use this Message. Electronic messages are not secure or error free and can contain viruses or may be delayed and the sender is not liable for any of these occurrences. The sender reserves the right to monitor, record, transfer cross border and retain electronic messages. "D&B" is a trading style of D&B Business Information Solutions is registered in Ireland.www.dnb.co.uk
Re: Run Yarn/Hadoop jar without first unjarring to /tmp folder?
Posted by Annabel Melongo <me...@yahoo.com.INVALID>.
O'Neill,
Yes, you can submit your job with jars without unpacking and depending on your cluster environment, the jars will or will not be automatically copied. This topic was discussed in this mailing list about two weeks ago. Please read Andrew Or's explanations, on Tuesday Dec. 20th, 2015, on how the Spark framework accomplishes this:
Re: Can't submit job to stand alone cluster
| |
| | | | | | | |
| Re: Can't submit job to stand alone clusterLet me clarify a few things for everyone: There are three *cluster managers*: standalone, YARN, and Mesos. Each cluster manager can run in two *deploy modes*, clie... |
| |
| View on www.mail-archive.com | Preview by Yahoo |
| |
|
|
On Friday, January 8, 2016 2:44 PM, "O'Neill, Paul" <On...@dnb.com.INVALID> wrote:
<!--#yiv5999690184 _filtered #yiv5999690184 {font-family:"Cambria Math";panose-1:2 4 5 3 5 4 6 3 2 4;} _filtered #yiv5999690184 {font-family:Calibri;panose-1:2 15 5 2 2 2 4 3 2 4;} _filtered #yiv5999690184 {font-family:"Gill Sans MT";panose-1:2 11 5 2 2 1 4 2 2 3;}#yiv5999690184 #yiv5999690184 p.yiv5999690184MsoNormal, #yiv5999690184 li.yiv5999690184MsoNormal, #yiv5999690184 div.yiv5999690184MsoNormal {margin:0cm;margin-bottom:.0001pt;font-size:11.0pt;font-family:"Calibri", sans-serif;}#yiv5999690184 a:link, #yiv5999690184 span.yiv5999690184MsoHyperlink {color:#0563C1;text-decoration:underline;}#yiv5999690184 a:visited, #yiv5999690184 span.yiv5999690184MsoHyperlinkFollowed {color:#954F72;text-decoration:underline;}#yiv5999690184 span.yiv5999690184EmailStyle17 {font-family:"Calibri", sans-serif;color:windowtext;}#yiv5999690184 .yiv5999690184MsoChpDefault {font-family:"Calibri", sans-serif;} _filtered #yiv5999690184 {margin:72.0pt 72.0pt 72.0pt 72.0pt;}#yiv5999690184 div.yiv5999690184WordSection1 {}-->Hey, Is it possible to execute a jar without unjarring ro /tmp first? We have large Shaded jars are so are running out of inode space on /tmp. I can execute without unpacking via java –jar and adding the output from “yarn classpath” however I am missing all the options that are set via the YARN script. Is there anyway to do this other than manually duplicating the entries in yarn script? Also, is it possible to change where the jars are unpacked to without also changing the Hadoop.tmp.dir across all the data nodes? It seems like these should be unrelated properties? Cheers. Paul O’Neill Principal Engineer The Chase Building, 5th Floor Carmanhall Road, Sandyford, Dublin 18, Ireland Direct +353 1 256 6090 Mobile +353 86 835 4084 http://www.dnb.co.uk/ The information contained in this electronic message and any attachments (the "Message") is intended for one or more specific individuals or entities, and may be confidential, proprietary, privileged or otherwise protected by law. If you are not the intended recipient (or you are not authorised to receive for the recipient), please notify the sender immediately, delete this Message and do not disclose, distribute, or copy it to any third party or otherwise use this Message. Electronic messages are not secure or error free and can contain viruses or may be delayed and the sender is not liable for any of these occurrences. The sender reserves the right to monitor, record, transfer cross border and retain electronic messages. "D&B" is a trading style of D&B Business Information Solutions is registered in Ireland.www.dnb.co.uk
Re: Run Yarn/Hadoop jar without first unjarring to /tmp folder?
Posted by Annabel Melongo <me...@yahoo.com.INVALID>.
O'Neill,
Yes, you can submit your job with jars without unpacking and depending on your cluster environment, the jars will or will not be automatically copied. This topic was discussed in this mailing list about two weeks ago. Please read Andrew Or's explanations, on Tuesday Dec. 20th, 2015, on how the Spark framework accomplishes this:
Re: Can't submit job to stand alone cluster
| |
| | | | | | | |
| Re: Can't submit job to stand alone clusterLet me clarify a few things for everyone: There are three *cluster managers*: standalone, YARN, and Mesos. Each cluster manager can run in two *deploy modes*, clie... |
| |
| View on www.mail-archive.com | Preview by Yahoo |
| |
|
|
On Friday, January 8, 2016 2:44 PM, "O'Neill, Paul" <On...@dnb.com.INVALID> wrote:
<!--#yiv5999690184 _filtered #yiv5999690184 {font-family:"Cambria Math";panose-1:2 4 5 3 5 4 6 3 2 4;} _filtered #yiv5999690184 {font-family:Calibri;panose-1:2 15 5 2 2 2 4 3 2 4;} _filtered #yiv5999690184 {font-family:"Gill Sans MT";panose-1:2 11 5 2 2 1 4 2 2 3;}#yiv5999690184 #yiv5999690184 p.yiv5999690184MsoNormal, #yiv5999690184 li.yiv5999690184MsoNormal, #yiv5999690184 div.yiv5999690184MsoNormal {margin:0cm;margin-bottom:.0001pt;font-size:11.0pt;font-family:"Calibri", sans-serif;}#yiv5999690184 a:link, #yiv5999690184 span.yiv5999690184MsoHyperlink {color:#0563C1;text-decoration:underline;}#yiv5999690184 a:visited, #yiv5999690184 span.yiv5999690184MsoHyperlinkFollowed {color:#954F72;text-decoration:underline;}#yiv5999690184 span.yiv5999690184EmailStyle17 {font-family:"Calibri", sans-serif;color:windowtext;}#yiv5999690184 .yiv5999690184MsoChpDefault {font-family:"Calibri", sans-serif;} _filtered #yiv5999690184 {margin:72.0pt 72.0pt 72.0pt 72.0pt;}#yiv5999690184 div.yiv5999690184WordSection1 {}-->Hey, Is it possible to execute a jar without unjarring ro /tmp first? We have large Shaded jars are so are running out of inode space on /tmp. I can execute without unpacking via java –jar and adding the output from “yarn classpath” however I am missing all the options that are set via the YARN script. Is there anyway to do this other than manually duplicating the entries in yarn script? Also, is it possible to change where the jars are unpacked to without also changing the Hadoop.tmp.dir across all the data nodes? It seems like these should be unrelated properties? Cheers. Paul O’Neill Principal Engineer The Chase Building, 5th Floor Carmanhall Road, Sandyford, Dublin 18, Ireland Direct +353 1 256 6090 Mobile +353 86 835 4084 http://www.dnb.co.uk/ The information contained in this electronic message and any attachments (the "Message") is intended for one or more specific individuals or entities, and may be confidential, proprietary, privileged or otherwise protected by law. If you are not the intended recipient (or you are not authorised to receive for the recipient), please notify the sender immediately, delete this Message and do not disclose, distribute, or copy it to any third party or otherwise use this Message. Electronic messages are not secure or error free and can contain viruses or may be delayed and the sender is not liable for any of these occurrences. The sender reserves the right to monitor, record, transfer cross border and retain electronic messages. "D&B" is a trading style of D&B Business Information Solutions is registered in Ireland.www.dnb.co.uk
Re: Run Yarn/Hadoop jar without first unjarring to /tmp folder?
Posted by Annabel Melongo <me...@yahoo.com.INVALID>.
O'Neill,
Yes, you can submit your job with jars without unpacking and depending on your cluster environment, the jars will or will not be automatically copied. This topic was discussed in this mailing list about two weeks ago. Please read Andrew Or's explanations, on Tuesday Dec. 20th, 2015, on how the Spark framework accomplishes this:
Re: Can't submit job to stand alone cluster
| |
| | | | | | | |
| Re: Can't submit job to stand alone clusterLet me clarify a few things for everyone: There are three *cluster managers*: standalone, YARN, and Mesos. Each cluster manager can run in two *deploy modes*, clie... |
| |
| View on www.mail-archive.com | Preview by Yahoo |
| |
|
|
On Friday, January 8, 2016 2:44 PM, "O'Neill, Paul" <On...@dnb.com.INVALID> wrote:
<!--#yiv5999690184 _filtered #yiv5999690184 {font-family:"Cambria Math";panose-1:2 4 5 3 5 4 6 3 2 4;} _filtered #yiv5999690184 {font-family:Calibri;panose-1:2 15 5 2 2 2 4 3 2 4;} _filtered #yiv5999690184 {font-family:"Gill Sans MT";panose-1:2 11 5 2 2 1 4 2 2 3;}#yiv5999690184 #yiv5999690184 p.yiv5999690184MsoNormal, #yiv5999690184 li.yiv5999690184MsoNormal, #yiv5999690184 div.yiv5999690184MsoNormal {margin:0cm;margin-bottom:.0001pt;font-size:11.0pt;font-family:"Calibri", sans-serif;}#yiv5999690184 a:link, #yiv5999690184 span.yiv5999690184MsoHyperlink {color:#0563C1;text-decoration:underline;}#yiv5999690184 a:visited, #yiv5999690184 span.yiv5999690184MsoHyperlinkFollowed {color:#954F72;text-decoration:underline;}#yiv5999690184 span.yiv5999690184EmailStyle17 {font-family:"Calibri", sans-serif;color:windowtext;}#yiv5999690184 .yiv5999690184MsoChpDefault {font-family:"Calibri", sans-serif;} _filtered #yiv5999690184 {margin:72.0pt 72.0pt 72.0pt 72.0pt;}#yiv5999690184 div.yiv5999690184WordSection1 {}-->Hey, Is it possible to execute a jar without unjarring ro /tmp first? We have large Shaded jars are so are running out of inode space on /tmp. I can execute without unpacking via java –jar and adding the output from “yarn classpath” however I am missing all the options that are set via the YARN script. Is there anyway to do this other than manually duplicating the entries in yarn script? Also, is it possible to change where the jars are unpacked to without also changing the Hadoop.tmp.dir across all the data nodes? It seems like these should be unrelated properties? Cheers. Paul O’Neill Principal Engineer The Chase Building, 5th Floor Carmanhall Road, Sandyford, Dublin 18, Ireland Direct +353 1 256 6090 Mobile +353 86 835 4084 http://www.dnb.co.uk/ The information contained in this electronic message and any attachments (the "Message") is intended for one or more specific individuals or entities, and may be confidential, proprietary, privileged or otherwise protected by law. If you are not the intended recipient (or you are not authorised to receive for the recipient), please notify the sender immediately, delete this Message and do not disclose, distribute, or copy it to any third party or otherwise use this Message. Electronic messages are not secure or error free and can contain viruses or may be delayed and the sender is not liable for any of these occurrences. The sender reserves the right to monitor, record, transfer cross border and retain electronic messages. "D&B" is a trading style of D&B Business Information Solutions is registered in Ireland.www.dnb.co.uk