Database Administrators Stack Exchange is a question and answer site for database professionals who wish to improve their database skills and learn from others in the community. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have begun writing a website using Membase (php memcached library) and I've come up against a conceptual wall. This is my scenario:

I have some data that is stored in a key-value pair which will be accessed and modified by more than one user. If one of these users wants to update the data then they:

1. Get the data and also get the CAS token.
2. Attempt to do a CAS write.
3. If the CAS write fails, start again from 1.
4. If the CAS write succeeds, move on.

What if I have two users trying to access the same piece of data? Thinking about this I can see that it is going to result in an infinite loop:

1. User 1 gets the data and also retrieves CAS token z1.
2. User 2 then gets the same data and now retrieves CAS token z2.
3. User 1 attempts to do a CAS write but it fails since z1 != z2
4. User 1 repeats step 1 and step 3 with user1's CAS tokens incrementing to z3, z5, z7, etc.
5. User 2 attempts to do a CAS write but it fails since z2 != z3.
6. User 2 repeats step 2 and step 5 with user 2's CAS tokens incrementing to z4, z6, z8, etc.

I know such a scenario is probably really uncommon since two users would have to try and modify the same piece of data at the same instant, but "uncommon" != "never".

In order to prevent this scenario I thought of introducing a random delay into the process so that steps 1-6 happen out of order and are therefore able to complete. However, I don't like this idea because it will slow the website down.

Is there a standard way of handling this type of problem?

I can give more info on the specific data I want to update if that helps, but this seems like a scenario I am likely to encounter quite a bit so I want to keep it general. One thing I will say though is that the process happens unbeknownst to both users - i.e. it is their browsers which are performing the updates, the action is not initiated by a human. Simply warning that 'an error occured - please try again' would not make sense since the users are not trying to do anything.

Please feel free to ask more questions if you need more detail.

edit after some more thought on this

I think I may have misunderstood the CAS token. I think the token only changes upon update so the above is not a problem. still, I'll leave it just in case anyone else has some wisdom to add.

share|improve this question

closed as too localized by Jack Douglas Sep 6 '12 at 9:45

This question is unlikely to help any future visitors; it is only relevant to a small geographic area, a specific moment in time, or an extraordinarily narrow situation that is not generally applicable to the worldwide audience of the internet. For help making this question more broadly applicable, visit the help center.If this question can be reworded to fit the rules in the help center, please edit the question.

considering your effort and the lack of response I think it's appropriate to close the question – Jack Douglas Sep 6 '12 at 9:45
yep. no wisdom added :P – mulllhausen Sep 7 '12 at 0:27
thanks for being good-humoured about it - I wish we had more NoSQL experts contributing who could add their thoughts! – Jack Douglas Sep 7 '12 at 6:06

Browse other questions tagged or ask your own question.