On Thu, Oct 13, 2016 at 4:44 PM, Tom Lane wrote:
> I was about to suggest that maybe we didn't need cleanup locks in btree
> indexes anymore now that SnapshotNow is gone, but I see that somebody
> already altered nbtree/README to say this:
>
> : Therefore, a scan using an MVCC snapshot which has
On Fri, Oct 14, 2016 at 2:52 AM, Robert Haas wrote:
> On Thu, Oct 13, 2016 at 6:33 AM, Amit Kapila wrote:
>> If we agree that above is a problematic case, then some of the options
>> to solve it could be (a) Vacuum should not wait for a cleanup lock and
>> instead just give up and start again whi
On Fri, Oct 14, 2016 at 3:14 AM, Tom Lane wrote:
> Robert Haas writes:
>> On Thu, Oct 13, 2016 at 6:33 AM, Amit Kapila wrote:
>>> If we agree that above is a problematic case, then some of the options
>>> to solve it could be (a) Vacuum should not wait for a cleanup lock and
>>> instead just giv
Robert Haas writes:
> On Thu, Oct 13, 2016 at 6:33 AM, Amit Kapila wrote:
>> If we agree that above is a problematic case, then some of the options
>> to solve it could be (a) Vacuum should not wait for a cleanup lock and
>> instead just give up and start again which I think is a bad idea (b)
>>
On Thu, Oct 13, 2016 at 6:33 AM, Amit Kapila wrote:
> If we agree that above is a problematic case, then some of the options
> to solve it could be (a) Vacuum should not wait for a cleanup lock and
> instead just give up and start again which I think is a bad idea (b)
> don't allow to take lock of
During Vacuum, cleanup in btree need to acquire cleanup lock on the
block and the non-MVCC scans or scans on unlogged index retain the pin
on a buffer during scan of a block. Now, if the scan is suspended
(use cursor and then fetch one row) after holding a pin on a buffer in
Session-1 and in Sessi