[
https://issues.apache.org/jira/browse/SPARK-3948?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14172371#comment-14172371
]
Saisai Shao commented on SPARK-3948:
------------------------------------
Hi [~mridulm80], thanks a lot for your suggestions, I tested with different
Java version and two OS: Redhat 6.2 and Ubuntu 14.04, I think this behavior is
OS or kernel related, not jvm version, you can see the output as below:
First is the result under Redhat, I tried Java 1.8.0_20, 1.7.0_60 and 1.7.0_04,
the result is:
{noformat}
size = 0
position = 0
size = 118 for i = 0
size = 118
position = 0
size = 118 for i = 1
size = 118
position = 0
size = 118 for i = 2
size = 118
position = 0
size = 118 for i = 3
size = 118
position = 0
size = 118 for i = 4
size = 118
position = 0
size = 118 for i = 5
size = 118
position = 0
size = 118 for i = 6
size = 118
position = 0
size = 118 for i = 7
size = 118
position = 0
size = 118 for i = 8
size = 118
position = 0
size = 118 for i = 9
{noformat}
Obvious the position is always 0, so the final output file size is just 118.
Then changed to Ubuntu 14.04, with Java 1.7.0_04 and 1.7.0_51, the result shows
as below:
{noformat}
size = 0
position = 0
size = 118 for i = 0
size = 118
position = 118
size = 118 for i = 1
size = 236
position = 236
size = 118 for i = 2
size = 354
position = 354
size = 118 for i = 3
size = 472
position = 472
size = 118 for i = 4
size = 590
position = 590
size = 118 for i = 5
size = 708
position = 708
size = 118 for i = 6
size = 826
position = 826
size = 118 for i = 7
size = 944
position = 944
size = 118 for i = 8
size = 1062
position = 1062
size = 118 for i = 9
{noformat}
This is the result we expected, and position also seek to the right end, so the
final output file size is 1180 as expected.
My Ubuntu machine's kernel version is: 3.13.0-37-generic; and Redhat machine's
kernel version is: 2.6.32-220.el6.x86_64.
So I guess maybe the difference relies on kernel version or OS version, but
still need to verify.
Besides I think to keep the consistency, we can add *append=true* flag to
enforce the same output even in different platforms.
> Sort-based shuffle can lead to assorted stream-corruption exceptions
> --------------------------------------------------------------------
>
> Key: SPARK-3948
> URL: https://issues.apache.org/jira/browse/SPARK-3948
> Project: Spark
> Issue Type: Bug
> Components: Shuffle
> Affects Versions: 1.2.0
> Reporter: Saisai Shao
> Assignee: Saisai Shao
>
> Several exceptions occurred when running TPC-DS queries against latest master
> branch with sort-based shuffle enable, like PARSING_ERROR(2) in snappy,
> deserializing error in Kryo and offset out-range in FileManagedBuffer, all
> these exceptions are gone when we changed to hash-based shuffle.
> With deep investigation, we found that some shuffle output file is
> unexpectedly smaller than the others, as the log shows:
> {noformat}
> 14/10/14 18:25:06 INFO shuffle.IndexShuffleBlockManager: Block id:
> shuffle_6_9_11, offset: 3055635, length: 236708, file length: 47274167
> 14/10/14 18:25:06 INFO shuffle.IndexShuffleBlockManager: Block id:
> shuffle_6_10_11, offset: 2986484, length: 222755, file length: 47174539
> 14/10/14 18:25:06 INFO shuffle.IndexShuffleBlockManager: Block id:
> shuffle_6_11_11, offset: 2995341, length: 259871, file length: 383405
> 14/10/14 18:25:06 INFO shuffle.IndexShuffleBlockManager: Block id:
> shuffle_6_12_11, offset: 2991030, length: 268191, file length: 47478892
> 14/10/14 18:25:06 INFO shuffle.IndexShuffleBlockManager: Block id:
> shuffle_6_13_11, offset: 3016292, length: 230694, file length: 47420826
> 14/10/14 18:25:06 INFO shuffle.IndexShuffleBlockManager: Block id:
> shuffle_6_14_11, offset: 3061400, length: 241136, file length: 47395509
> {noformat}
> As you can see the total file length of shuffle_6_11_11 is much smaller than
> other same stage map output results.
> And we also dump the map outputs in map side to see if this small size output
> is correct or not, below is the log:
> {noformat}
> In bypass merge sort, file name: /mnt/DP_disk1/animal/spark/spark-local-
> 20141014182142-8345/22/shuffle_6_11_0.data, file length: 383405length:
> 274722 262597 291290 272902 264941 270358 291005 295285 252482
> 287142 232617 259871 233734 241439 228897 234282 253834 235619
> 233803 255532 270739 253825 262087 266404 234273 250120 262983
> 257024 255947 254971 258908 247862 221613 258566 245399 251684
> 274843 226150 264278 245279 225656 235084 239466 212851 242245
> 218781 222191 215500 211548 234256 208601 204113 191923 217895
> 227020 215331 212313 223725 250876 256875 239276 266777 235520
> 237462 234063 242270 246825 255888 235937 236956 233099 264508
> 260303 233294 239061 254856 257475 230105 246553 260412 210355
> 211201 219572 206636 226866 209937 226618 218208 206255 248069
> 221717 222112 215734 248088 239207 246125 239056 241133 253091
> 246738 233128 242794 231606 255737 221123 252115 247286 229688
> 251087 250047 237579 263079 256251 238214 208641 201120 204009
> 200825 211965 200600 194492 226471 194887 226975 215072 206008
> 233288 222132 208860 219064 218162 237126 220465 201343 225711
> 232178 233786 212767 211462 213671 215853 227822 233782 214727
> 247001 228968 247413 222674 214241 184122 215643 207665 219079
> 215185 207718 212723 201613 216600 212591 208174 204195 208099
> 229079 230274 223373 214999 256626 228895 231821 383405 229646
> 220212 245495 245960 227556 213266 237203 203805 240509 239306
> 242365 218416 238487 219397 240026 251011 258369 255365 259811
> 283313 248450 264286 264562 257485 279459 249187 257609 274964
> 292369 273826
> {noformat}
> Here I dump the file name, length and each partition's length, obviously the
> sum of all partition lengths is not equal to file length. So I think there
> may be a situation paritionWriter in ExternalSorter not always append to the
> end of previous written file, the file's content is overwritten in some
> parts, and this lead to the exceptions I mentioned before.
> Also I changed the code of copyStream by disable transferTo, use the previous
> one, all the issues are gone. So I think there maybe some flushing problems
> in transferTo when processed data is large.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
