[
https://issues.apache.org/jira/browse/BOOKKEEPER-11?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13055515#comment-13055515
]
Ivan Kelly commented on BOOKKEEPER-11:
--------------------------------------
{quote}
The problem is that the writer of the ledger might have failed at the time the
application invokes "asyncOpenActiveLedger". Consequently, the ledger might not
really be active anymore, and the application must know of that possibility. I
don't mind changing the name, but I don't feel that OpenActive makes it better.
{quote}
This is true. "Active" doesn't adequately describe either given this. How about
openLedgerReadOnly? I have doubts about this too because usually opening a
ledger is a read only op anyhow.
{quote}
Ok. Listing the variables will help.
{quote}
So, in this patch, LedgerOpenOp#unsafe, ReadLastConfirmedOp#lh &
ReadLastConfirmedOp#ctx are the new member variables which should be final.
{quote}
Due to concurrent writes and striping, we are not guaranteed to have a quorum
returning the same last hint. But, given that we compute a digest and verify
upon reading, whatever value we return should be correct. In fact, it is ok to
return the first value we receive. I've made it wait for a quorum because by
assumption there must be at least one quorum available, ans we have a chance of
returning a larger value (assuming there is one) if we read from more bookies .
Waiting for more than a quorum does not guarantee that the operation will
return.
{quote}
Ah, I think my understanding was that I didnt see it was using lastAddConfirmed
and not entryId itself.
> Read from open ledger
> ---------------------
>
> Key: BOOKKEEPER-11
> URL: https://issues.apache.org/jira/browse/BOOKKEEPER-11
> Project: Bookkeeper
> Issue Type: New Feature
> Reporter: Flavio Junqueira
> Attachments: BOOKKEEPER-11.patch, zk-1001-design-doc.pdf,
> zk-1001-design-doc.pdf
>
>
> The BookKeeper client currently does not allow a client to read from an open
> ledger. That is, if the creator of a ledger is still writing to it (and the
> ledger is not closed), then an attempt to open the same ledger for reading
> will execute the code to recover the ledger, assuming that the ledger has not
> been correctly closed.
> It seems that there are applications that do require the ability to read from
> a ledger while it is being written to, and the main goal of this jira is to
> discuss possible implementations of this feature.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
