Indexes completely erased

I recently reinstalled my OS (upgrading from Ubuntu 32-bit to 64-bit)
but left the storage containing ES untouched. Upon restarting ES I
saw several errors in the log. I was on a fresh install and the log
seemed to indicate I had not changed the nofiles limits. After
changing this to 32000 and restarting there were no more errors in the
log however I was horrified to see that a good chunk of my indexes had
been completely erased (that is they still existed by were now
completely empty )! A gist of a small portion of my log file can be
found at :

https://gist.github.com/938449

-brian

When did you build the 0.16 version? How many nodes were in the cluster?
On Saturday, April 23, 2011 at 10:46 AM, merrellb wrote:

I recently reinstalled my OS (upgrading from Ubuntu 32-bit to 64-bit)
but left the storage containing ES untouched. Upon restarting ES I
saw several errors in the log. I was on a fresh install and the log
seemed to indicate I had not changed the nofiles limits. After
changing this to 32000 and restarting there were no more errors in the
log however I was horrified to see that a good chunk of my indexes had
been completely erased (that is they still existed by were now
completely empty )! A gist of a small portion of my log file can be
found at :

https://gist.github.com/938449

-brian

It appears it was build from Master about a month ago (March 16th). I
am only running a single node.

On Apr 23, 6:14 am, Shay Banon shay.ba...@elasticsearch.com wrote:

When did you build the 0.16 version? How many nodes were in the cluster?

On Saturday, April 23, 2011 at 10:46 AM, merrellb wrote:

I recently reinstalled my OS (upgrading from Ubuntu 32-bit to 64-bit)
but left the storage containing ES untouched. Upon restarting ES I
saw several errors in the log. I was on a fresh install and the log
seemed to indicate I had not changed the nofiles limits. After
changing this to 32000 and restarting there were no more errors in the
log however I was horrified to see that a good chunk of my indexes had
been completely erased (that is they still existed by were now
completely empty )! A gist of a small portion of my log file can be
found at :

https://gist.github.com/938449

-brian

I think, based on the failure, that this was fixed in latest master (for a single node case). 0.16 will be released today. Sorry about that!
On Saturday, April 23, 2011 at 6:04 PM, merrellb wrote:

It appears it was build from Master about a month ago (March 16th). I
am only running a single node.

On Apr 23, 6:14 am, Shay Banon shay.ba...@elasticsearch.com wrote:

When did you build the 0.16 version? How many nodes were in the cluster?

On Saturday, April 23, 2011 at 10:46 AM, merrellb wrote:

I recently reinstalled my OS (upgrading from Ubuntu 32-bit to 64-bit)
but left the storage containing ES untouched. Upon restarting ES I
saw several errors in the log. I was on a fresh install and the log
seemed to indicate I had not changed the nofiles limits. After
changing this to 32000 and restarting there were no more errors in the
log however I was horrified to see that a good chunk of my indexes had
been completely erased (that is they still existed by were now
completely empty )! A gist of a small portion of my log file can be
found at :

https://gist.github.com/938449

-brian

Thanks Shay. Certainly not fun losing data although I suppose such
are the risks of running Master. Glad the fix was able to make it
into 0.16.

-brian

On Apr 23, 4:25 pm, Shay Banon shay.ba...@elasticsearch.com wrote:

I think, based on the failure, that this was fixed in latest master (for a single node case). 0.16 will be released today. Sorry about that!

On Saturday, April 23, 2011 at 6:04 PM, merrellb wrote:

It appears it was build from Master about a month ago (March 16th). I
am only running a single node.

On Apr 23, 6:14 am, Shay Banon shay.ba...@elasticsearch.com wrote:

When did you build the 0.16 version? How many nodes were in the cluster?

On Saturday, April 23, 2011 at 10:46 AM, merrellb wrote:

I recently reinstalled my OS (upgrading from Ubuntu 32-bit to 64-bit)
but left the storage containing ES untouched. Upon restarting ES I
saw several errors in the log. I was on a fresh install and the log
seemed to indicate I had not changed the nofiles limits. After
changing this to 32000 and restarting there were no more errors in the
log however I was horrified to see that a good chunk of my indexes had
been completely erased (that is they still existed by were now
completely empty )! A gist of a small portion of my log file can be
found at :

https://gist.github.com/938449

-brian

did you change the elasticsearch.json ?

On Apr 24, 6:35 am, merrellb br...@merrells.org wrote:

Thanks Shay. Certainly not fun losing data although I suppose such
are the risks of running Master. Glad the fix was able to make it
into 0.16.

-brian

On Apr 23, 4:25 pm, Shay Banon shay.ba...@elasticsearch.com wrote:

I think, based on the failure, that this was fixed in latest master (for a single node case). 0.16 will be released today. Sorry about that!

On Saturday, April 23, 2011 at 6:04 PM, merrellb wrote:

It appears it was build from Master about a month ago (March 16th). I
am only running a single node.

On Apr 23, 6:14 am, Shay Banon shay.ba...@elasticsearch.com wrote:

When did you build the 0.16 version? How many nodes were in the cluster?

On Saturday, April 23, 2011 at 10:46 AM, merrellb wrote:

I recently reinstalled my OS (upgrading from Ubuntu 32-bit to 64-bit)
but left the storage containing ES untouched. Upon restarting ES I
saw several errors in the log. I was on a fresh install and the log
seemed to indicate I had not changed the nofiles limits. After
changing this to 32000 and restarting there were no more errors in the
log however I was horrified to see that a good chunk of my indexes had
been completely erased (that is they still existed by were now
completely empty )! A gist of a small portion of my log file can be
found at :

https://gist.github.com/938449

-brian