Обсуждение: [ADMIN] postgresql : could not serialize access due to read/writedependencies among transactions

On Tue, Jan 17, 2017 at 10:54 PM, Neslisah Demirci
<neslisah.demirci@markafoni.com> wrote:

> could not serialize access due to read/write dependencies among
> transactions

> I also add an index to my query and my query's execution plan
> don't use seq scan .

These two issues are likely to be somewhat related -- if a
sequential scan is used, then any write to that table by another
connection causes a read-write dependency (a/k/a rw-conflict),
which can eventually contribute to a serialization failure.  If you
can cause narrower access through indexes, you may see a
significant drop in the frequency of these serialization failures.
You might want to post the query and its execution plan with all
the information suggested here (exact pg version, configuration
information, machine descriptions, etc.):

https://wiki.postgresql.org/wiki/SlowQueryQuestions

On the other hand, if you are going to use serializable
transactions (or even repeatable read transactions) you should
probably be using some framework that can retry the transaction
from the start on a serialization failure.

You might be interested in this set of examples of how serializable
transactions differ from repeatable read:

https://wiki.postgresql.org/wiki/SSI

And of course, if you haven't already read the fine manual on the
topic:

https://www.postgresql.org/docs/current/static/mvcc.html

--
Kevin Grittner
EDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

Am 01/18/2017 um 05:54 AM schrieb Neslisah Demirci:
> Hi ,
>
>
> I have problem about  this issue ;
>
>
> could not serialize access due to read/write dependencies among transactions

That's something you have to expect when using SERIALIZABLE isolation
level. Maybe re-read
https://www.postgresql.org/docs/current/static/transaction-iso.html#XACT-SERIALIZABLE


> A message app like whatsapp i decided to use isolation level
> serializable if i use other transaction levels many conversations
> started with same number (think about whatsapp web sidebar multipling
> with same number every message).

Are you not using a sequence for generating these IDs? As those are
counting independent of transaction visibility...
What you say sounds more like "SELECT max(conversation_id) + 1 AS
new_conversation_id FROM ..."

> I also add an index to my query and my query's execution plan don't use
> seq scan .
>
>
> How can i solve this issue ? Can i solve this on db orr on app  ?

a) switch to a sequence for generating these IDs (you can go back to a
lower isolation level then). You may get holes in the IDs then (on
rolled back transactions), but I don't see how that would not be
acceptable for conversation IDs
b) deal with it in the app (probably not what you want, think roud-trip
time)


> Neslişah Demirci | Veritabanı Yöneticisi
>
> Ayazağa cad. No:4 Uniq İstanbul Plaza
> B2 /Kat:3 34396 Ayazağa-SARIYER-İstanbul
> T. (+90) 212 453 16 00 – 5516
> F. (+90) 212 453 16 16
> www.markafoni.com <http://www.markafoni.com/>
> www.facebook.com/markafoni <http://www.facebook.com/markafoni>
> blog.markafoni.com
>
>
>

Regards,
--
Gunnar "Nick" Bluth
DBA ELSTER

Tel:   +49 911/991-4665
Mobil: +49 172/8853339