FAQ
HI all,

I updated from 0.6 to 0.7.1 using parcel feature in Cloudera Manager 4.5.

When I run even the very simple query (select * from TABLE limit 1),
every impalad daemon on nodes always crashes.
impala-shell says "Error communicating with impalad: TSocket read 0 bytes".

And impalad daemon also outputs log file of "hs_err_pidXXXX.log" (maybe
XXXX is process id).
"A fatal error has been detected by the Java Runtime Environment:" is
written in the head of this file.
Is this problem of Java ?

Of course, the same query on both hive and (rollbacked) impala 0.6 works
fine.

Has anyone seen similar behavior ?

The attached file is log file above.

Thanks,

suda

Search Discussions

  • Yukinori SUDA at Apr 19, 2013 at 6:56 am
    Hi 邱睿,

    thank you for your report.

    I think that they result from the same reason.

    Thanks,

    suda

    2013/4/19 邱睿 <rayqiu0509@gmail.com>
    I meet the exactly same problem. Impalad daemon always crashes on
    following frame:
    C [impalad+0x4ed2a7] _ZN6impala12HdfsScanNode14UpdateCountersEv+0x97

    在 2013年4月19日星期五UTC+8下午1时16分09秒,Suda Yukinori写道:
    HI all,

    I updated from 0.6 to 0.7.1 using parcel feature in Cloudera Manager 4.5.

    When I run even the very simple query (select * from TABLE limit 1),
    every impalad daemon on nodes always crashes.
    impala-shell says "Error communicating with impalad: TSocket read 0
    bytes".

    And impalad daemon also outputs log file of "hs_err_pidXXXX.log" (maybe
    XXXX is process id).
    "A fatal error has been detected by the Java Runtime Environment:" is
    written in the head of this file.
    Is this problem of Java ?

    Of course, the same query on both hive and (rollbacked) impala 0.6 works
    fine.

    Has anyone seen similar behavior ?

    The attached file is log file above.

    Thanks,

    suda
  • Henry Robinson at Apr 19, 2013 at 7:41 am
    Thanks for your bug reports, and apologies for this issue, which I've
    filed https://issues.cloudera.org/browse/IMPALA-289 to track. We may be
    asking for more details shortly, but we hope to have it fixed soon.

    Best,
    Henry


    On 18 April 2013 23:56, Yukinori SUDA wrote:

    Hi 邱睿,

    thank you for your report.

    I think that they result from the same reason.

    Thanks,

    suda

    2013/4/19 邱睿 <rayqiu0509@gmail.com>
    I meet the exactly same problem. Impalad daemon always crashes on
    following frame:
    C [impalad+0x4ed2a7] _ZN6impala12HdfsScanNode14UpdateCountersEv+0x97

    在 2013年4月19日星期五UTC+8下午1时16分09秒,Suda Yukinori写道:
    HI all,

    I updated from 0.6 to 0.7.1 using parcel feature in Cloudera Manager 4.5.

    When I run even the very simple query (select * from TABLE limit 1),
    every impalad daemon on nodes always crashes.
    impala-shell says "Error communicating with impalad: TSocket read 0
    bytes".

    And impalad daemon also outputs log file of "hs_err_pidXXXX.log" (maybe
    XXXX is process id).
    "A fatal error has been detected by the Java Runtime Environment:" is
    written in the head of this file.
    Is this problem of Java ?

    Of course, the same query on both hive and (rollbacked) impala 0.6 works
    fine.

    Has anyone seen similar behavior ?

    The attached file is log file above.

    Thanks,

    suda

    --
    Henry Robinson
    Software Engineer
    Cloudera
    415-994-6679
  • Yukinori SUDA at Apr 19, 2013 at 8:21 am
    Hi Henry,

    thank you for your creating issue.
    I'll watch this issue on JIRA.
    If you have any further questions, please feel free to contact me.

    I'm looking forward to when impala GA is released very much !!

    Thanks,

    suda

    2013/4/19 Henry Robinson <henry@cloudera.com>
    Thanks for your bug reports, and apologies for this issue, which I've
    filed https://issues.cloudera.org/browse/IMPALA-289 to track. We may be
    asking for more details shortly, but we hope to have it fixed soon.

    Best,
    Henry


    On 18 April 2013 23:56, Yukinori SUDA wrote:

    Hi 邱睿,

    thank you for your report.

    I think that they result from the same reason.

    Thanks,

    suda

    2013/4/19 邱睿 <rayqiu0509@gmail.com>
    I meet the exactly same problem. Impalad daemon always crashes on
    following frame:
    C [impalad+0x4ed2a7] _ZN6impala12HdfsScanNode14UpdateCountersEv+0x97

    在 2013年4月19日星期五UTC+8下午1时16分09秒,Suda Yukinori写道:
    HI all,

    I updated from 0.6 to 0.7.1 using parcel feature in Cloudera Manager
    4.5.

    When I run even the very simple query (select * from TABLE limit 1),
    every impalad daemon on nodes always crashes.
    impala-shell says "Error communicating with impalad: TSocket read 0
    bytes".

    And impalad daemon also outputs log file of "hs_err_pidXXXX.log" (maybe
    XXXX is process id).
    "A fatal error has been detected by the Java Runtime Environment:" is
    written in the head of this file.
    Is this problem of Java ?

    Of course, the same query on both hive and (rollbacked) impala 0.6
    works fine.

    Has anyone seen similar behavior ?

    The attached file is log file above.

    Thanks,

    suda

    --
    Henry Robinson
    Software Engineer
    Cloudera
    415-994-6679

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupimpala-user @
categorieshadoop
postedApr 19, '13 at 5:16a
activeApr 19, '13 at 8:21a
posts4
users2
websitecloudera.com
irc#hadoop

2 users in discussion

Yukinori SUDA: 3 posts Henry Robinson: 1 post

People

Translate

site design / logo © 2022 Grokbase