Steven ,
Do you have stacks trace for these exceptions ?

What kind of operations do you perform, I mean is it possible that you very
frequently change the same record again and again ?

Could we do following you wait for next issue and then describe in details
circumstances under which it happened ?
So we will have exception , operations which were performed or usually
performed on this kind of data and last event which lead to this issue.
Also I do suggest to update to 1.7.3 there's issue which is related to
server durability.

Also couple of words as we test cluster durability. There are 2 approaches .
1. We do thousands of operations to cover all branches of cluster code .
Then we copy wal , restore cluster content from scratch and compare byte by
bite content of files of original and restored cluster and they are
identical.
2. Crash the server during of parallel insertion of records for example and
check how much records were restored after server restart.

Also as I wrote before that other clients experienced server crash but
always managed to restore data . So your use case should be different in
some details and we will be really appreciate if you help us understand why
and looking forward to communicate with you to uncover all details of this
problem.
13 июня 2014 г. 0:05 пользователь "StevenTomer" <[email protected]>
написал:

> Andrey,
>
> I wish that were possible, but the customer data is confidential,
> making that kind of troubleshooting impossible.
>
> Steve
>
>
>
> On Thursday, June 12, 2014 1:19:13 PM UTC-6, Andrey Lomakin wrote:
>>
>> Steven,
>> We have clients which use different versions of rc of 1.7 for a months
>> and the only issue they experienced was Snappy serialization problem and we
>> did not see issues with clusters if records in cluster are broken.
>> So these issues are very strange.
>>
>> Any way, is it possible to access problematic database without sending it
>> by ssh for example ?
>> Or for example you connect us and we will write you where to look to
>> check what is wrong ?
>>
>>  --
>
> ---
> You received this message because you are subscribed to the Google Groups
> "OrientDB" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> For more options, visit https://groups.google.com/d/optout.
>

-- 

--- 
You received this message because you are subscribed to the Google Groups 
"OrientDB" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to