hadoop-common-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Steve Loughran <ste...@apache.org>
Subject Re: critical name node problem
Date Mon, 08 Sep 2008 11:25:46 GMT
Allen Wittenauer wrote:
> On 9/5/08 5:53 AM, "Andreas Kostyrka" <andreas@kostyrka.org> wrote:
>> Another idea would be a tool or namenode startup mode that would make it
>> ignore EOFExceptions to recover as much of the edits as possible.
>     We clearly need to change the "how to configure" docs to make sure
> people put at least two directories on two different storage systems for the
> dfs.name.dir  .  This problem seems to happen quite often, and having two+
> dirs helps protect against it.
>     We recently had one of the disks on one of our copies go bad.  The
> system kept going just fine until we had a chance to reconfig the name node.
>     That said, I've just HADOOP-4080 to help alert admins in these
> situations.

that and HADOOP-4081.

Apache Axis has this production/development switch; in develop mode it 
sends stack traces over the wire and is generally more forgiving. By 
default it assumes you are in production rather than development, so you 
have to explicitly flip the switch to get slighly reduced security.

Hadoop could have something similar, where if the  production flag is 
set, the cluster would simply refuse to come up if it felt the 
configuration wasn't robust enough.

View raw message