Originally the problem server was mtab2:

http://pastebin.com/8BpwBQQR

But, none of the rest servers could open it.  After restarting mtab2,
it got reassigned, and started to serve.

-Jack

On Mon, Oct 4, 2010 at 9:28 PM, Stack <[email protected]> wrote:
> It looks like it was just taking a while to open.  Is that possible?
>
> See there are a bunch of this:
>
> "010-10-04 19:30:57,557 INFO
> org.apache.hadoop.hbase.master.ServerManager: Processing
> MSG_REPORT_PROCESS_OPEN:
> img814,emmyrossum453.png,1284954679871.603680898 from
> mtab3.prod.imageshack.com,60020,1286241258709; 3 of 7"
>
> Thats code for "I'm working on it" w/ the usual holdup being a bunch
> of regions ahead of this one in the queue-to-open.
>
> You could take a look in the logs of mtab3.prod.imageshack.com from
> around this time to confirm.
>
> Then at the end of your log, I see this:
>
> 2010-10-04 19:30:58,537 INFO
> org.apache.hadoop.hbase.master.RegionServerOperation:
> img814,emmyrossum453.png,1284954679871.603680898 open on
> mtab3.prod.imageshack.com,60020,1286241258709
>
> Which usually means successful open.  Can you grab a row out of that region?
>
> St.Ack
>
> On Mon, Oct 4, 2010 at 8:54 PM, Jack Levin <[email protected]> wrote:
>> 0.89 0830, http://pastebin.com/y6m3ySNi
>>
>> Yes, I can hold on, its not super urgent, the previous email about the
>> unload regions is more troublesome.
>>
>> -Jack
>>
>> On Mon, Oct 4, 2010 at 8:43 PM, Stack <[email protected]> wrote:
>>> Can you grep mg814,emmyrossum453.png,1284954679871 in master log and
>>> paste the master log from when it starts showing up?
>>>
>>> You are on 0.89.x, right? (0724?)
>>>
>>> The assignment code and balancer are completely different in TRUNK,
>>> even different from the RC J-D is about to post.  Can you hang on a
>>> while or do you want to figure the issue?
>>>
>>> Thanks Jack,
>>> St.Ack
>>>
>>> On Mon, Oct 4, 2010 at 7:36 PM, Jack Levin <[email protected]> wrote:
>>>> http://pastebin.com/3fwiSkTF
>>>>
>>>> I kept getting this exception for a region until I restarted the
>>>> affected server, any ideas why master did not reassign it?
>>>>
>>>> -Jack
>>>>
>>>
>>
>

Reply via email to