My Hive QA job failed with "No space left on device (28)"

[Matthew McCline <mm...@hortonworks.com>]

https://builds.apache.org/job/PreCommit-HIVE-Build/10785?


Thanks

Re: My Hive QA job failed with "No space left on device (28)"

[Vihang Karajgaonkar <vi...@cloudera.com>]

I am seeing a lot of statements in hive.log which look like these

2018-05-07T05:28:15,113 DEBUG [main] ldif.LdifReader: Ldif version :
12018-05-07T05:28:15,114
DEBUG [main] ldif.LdifReader: Read an entry : dn:
m-oid=2.5.13.19,ou=normalizers,cn=system,ou=schemachangetype:
addcreatetimestamp:
20090818022727Zentrycsn: 20090818052727.840000Z#000000#000#000000m-oid:
2.5.13.19m-fqcn:
org.apache.directory.shared.ldap.schema.normalizers.NoOpNormalizercreatorsname:
uid=admin,ou=systemobjectclass: metaNormalizerobjectclass: metaTopobjectclass:
top
Does anyone know what these are and why they are spewing at such a high
frequency?

On Wed, May 9, 2018 at 4:05 PM, Vihang Karajgaonkar <vi...@cloudera.com>
wrote:

> Adding the correct dev list.
>
> It looks like we are generating a lot more logs than usual. Each precommit
> is generating ~22G of logs and the disk eventually ran out. I had to delete
> logs for PreCommit-HIVE-Build-1075[0-9] and
> PreCommit-HIVE-Build-1074[5-9] jobs to release the disk space. This usually
> happens if some patch adds a super-verbose debug log.
>
>
> On Wed, May 9, 2018 at 3:41 PM, Vihang Karajgaonkar <vi...@cloudera.com>
> wrote:
>
>> Hi Matt,
>>
>> Thanks for reporting the issue. Where do you see this error message? I
>> tried to find it but couldn't search for this in the console output.
>>
>> Thanks,
>> Vihang
>>
>> On Wed, May 9, 2018 at 2:05 PM, Matthew McCline <mmccline@hortonworks.com
>> > wrote:
>>
>>> https://builds.apache.org/job/PreCommit-HIVE-Build/10785​
>>>
>>>
>>> Thanks
>>>
>>
>>
>

Re: My Hive QA job failed with "No space left on device (28)"

[Vihang Karajgaonkar <vi...@cloudera.com>]

Adding the correct dev list.

It looks like we are generating a lot more logs than usual. Each precommit
is generating ~22G of logs and the disk eventually ran out. I had to delete
logs for PreCommit-HIVE-Build-1075[0-9] and PreCommit-HIVE-Build-1074[5-9]
jobs to release the disk space. This usually happens if some patch adds a
super-verbose debug log.


On Wed, May 9, 2018 at 3:41 PM, Vihang Karajgaonkar <vi...@cloudera.com>
wrote:

> Hi Matt,
>
> Thanks for reporting the issue. Where do you see this error message? I
> tried to find it but couldn't search for this in the console output.
>
> Thanks,
> Vihang
>
> On Wed, May 9, 2018 at 2:05 PM, Matthew McCline <mm...@hortonworks.com>
> wrote:
>
>> https://builds.apache.org/job/PreCommit-HIVE-Build/10785​
>>
>>
>> Thanks
>>
>
>