Reads

[Firas Khasawneh]

Hi,

 

I am trying to use Omid to support transactions in an HBase datastore. Is it possible to use Omid only for writes? Or do I need to use it for reads also? The reason I am asking is that I am trying to use Apache Drill and Apache Phoenix and if I use Omid for reads, this means I have to use the raw APIs of HBase and I would not be able to utilize Apache Drill or Phoenix sql layer. Please reply to me directly as I am not part of this group yet.

 

Thanks,

Firas Khasawneh

[Maysam Yabandeh]

Hi Firas,

In short, yes, you do need to wrap your reads with transactions as well. The reason is that after writes are performed there will be potentially multiple versions for each cell in Hbase (some aborted, some committed, and some from in-progress transactions)  and you would need transactional read to figure which version should be used.

Cheers

Mausa,

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

[Firas Khasawneh]

Hi Maysam,

 

Thanks for your response. What if I use the max timestamp per the paragraph below to read the record?

 

·         “Single-row Read. Each read in transaction txn_r_ must observe the last committed data before Ts(txn_r_).

·           To do so, starting with the latest version (assuming that the versions are sorted by timestamp in ascending order),

·         it looks for the first value with commit timestamp δ, where δ < Ts(txn_r_).

·         To verify, the transaction inquires the txn_w_ commit time of the TSO or its local, read-only replica on the client side.”

 

Thanks,

Firas