You are viewing a plain text version of this content. The canonical link for it is here.
Posted to hdfs-dev@hadoop.apache.org by Thanh Do <th...@cs.wisc.edu> on 2011/04/15 00:55:51 UTC

silent data loss during append

Hi all,

I have recently seen silent data loss in our system.
Here is the case:

1. client appends to some block
2. for some reason, commitBlockSynchronization
    returns successfully with synclist = [] (i.e empty)
3. in the client code, NO exception is thrown, and
   client appends successfully.
4. However, the block replicas are then removed from
   datanodes, causing data loss.

Have any one seen this before?
Is this behavior by design or a bug?

Many thanks,
Thanh

Re: silent data loss during append

Posted by "Tsz Wo (Nicholas), Sze" <s2...@yahoo.com>.
Hi Thanh,

There were some known bugs in the append feature in Apache Hadoop 0.19.  These 
bugs were fixed in both Apache 0.20-append and Apache 0.21.  For Cloudera's 
distribution, I have no idea.  You may want to ask your questions in Cloudera's 
mailing lists.

I am sorry for the bugs.

Regards,
Nicholas





________________________________
From: Thanh Do <th...@cs.wisc.edu>
To: Ted Dunning <td...@maprtech.com>
Cc: hdfs-user@hadoop.apache.org; hdfs-dev@hadoop.apache.org
Sent: Fri, April 15, 2011 8:02:05 AM
Subject: Re: silent data loss during append

I am using cloudera's distribution version:
hadoop-0.20.2+738.

On Thu, Apr 14, 2011 at 6:23 PM, Ted Dunning <td...@maprtech.com> wrote:

> What version are you using?
>
>
> On Thu, Apr 14, 2011 at 3:55 PM, Thanh Do <th...@cs.wisc.edu> wrote:
>
>> Hi all,
>>
>> I have recently seen silent data loss in our system.
>> Here is the case:
>>
>> 1. client appends to some block
>> 2. for some reason, commitBlockSynchronization
>>     returns successfully with synclist = [] (i.e empty)
>> 3. in the client code, NO exception is thrown, and
>>    client appends successfully.
>> 4. However, the block replicas are then removed from
>>    datanodes, causing data loss.
>>
>> Have any one seen this before?
>> Is this behavior by design or a bug?
>>
>> Many thanks,
>> Thanh

Re: silent data loss during append

Posted by Thanh Do <th...@cs.wisc.edu>.
I am using cloudera's distribution version:
hadoop-0.20.2+738.

On Thu, Apr 14, 2011 at 6:23 PM, Ted Dunning <td...@maprtech.com> wrote:

> What version are you using?
>
>
> On Thu, Apr 14, 2011 at 3:55 PM, Thanh Do <th...@cs.wisc.edu> wrote:
>
>> Hi all,
>>
>> I have recently seen silent data loss in our system.
>> Here is the case:
>>
>> 1. client appends to some block
>> 2. for some reason, commitBlockSynchronization
>>     returns successfully with synclist = [] (i.e empty)
>> 3. in the client code, NO exception is thrown, and
>>    client appends successfully.
>> 4. However, the block replicas are then removed from
>>    datanodes, causing data loss.
>>
>> Have any one seen this before?
>> Is this behavior by design or a bug?
>>
>> Many thanks,
>> Thanh
>>
>>
>>
>>
>

Re: silent data loss during append

Posted by Thanh Do <th...@cs.wisc.edu>.
I am using cloudera's distribution version:
hadoop-0.20.2+738.

On Thu, Apr 14, 2011 at 6:23 PM, Ted Dunning <td...@maprtech.com> wrote:

> What version are you using?
>
>
> On Thu, Apr 14, 2011 at 3:55 PM, Thanh Do <th...@cs.wisc.edu> wrote:
>
>> Hi all,
>>
>> I have recently seen silent data loss in our system.
>> Here is the case:
>>
>> 1. client appends to some block
>> 2. for some reason, commitBlockSynchronization
>>     returns successfully with synclist = [] (i.e empty)
>> 3. in the client code, NO exception is thrown, and
>>    client appends successfully.
>> 4. However, the block replicas are then removed from
>>    datanodes, causing data loss.
>>
>> Have any one seen this before?
>> Is this behavior by design or a bug?
>>
>> Many thanks,
>> Thanh
>>
>>
>>
>>
>

Re: silent data loss during append

Posted by Ted Dunning <td...@maprtech.com>.
What version are you using?

On Thu, Apr 14, 2011 at 3:55 PM, Thanh Do <th...@cs.wisc.edu> wrote:

> Hi all,
>
> I have recently seen silent data loss in our system.
> Here is the case:
>
> 1. client appends to some block
> 2. for some reason, commitBlockSynchronization
>     returns successfully with synclist = [] (i.e empty)
> 3. in the client code, NO exception is thrown, and
>    client appends successfully.
> 4. However, the block replicas are then removed from
>    datanodes, causing data loss.
>
> Have any one seen this before?
> Is this behavior by design or a bug?
>
> Many thanks,
> Thanh
>
>
>
>

Re: silent data loss during append

Posted by Ted Dunning <td...@maprtech.com>.
What version are you using?

On Thu, Apr 14, 2011 at 3:55 PM, Thanh Do <th...@cs.wisc.edu> wrote:

> Hi all,
>
> I have recently seen silent data loss in our system.
> Here is the case:
>
> 1. client appends to some block
> 2. for some reason, commitBlockSynchronization
>     returns successfully with synclist = [] (i.e empty)
> 3. in the client code, NO exception is thrown, and
>    client appends successfully.
> 4. However, the block replicas are then removed from
>    datanodes, causing data loss.
>
> Have any one seen this before?
> Is this behavior by design or a bug?
>
> Many thanks,
> Thanh
>
>
>
>