Re: logical decoding : exceeded maxAllocatedDescs for .spill files - Mailing list pgsql-hackers

From Amit Kapila
Subject Re: logical decoding : exceeded maxAllocatedDescs for .spill files
Date
Msg-id CAA4eK1LYw+BmANkXn2me_a3saMH_THGwBLSgrO_QB6u4ytLsmw@mail.gmail.com
Whole thread Raw
In response to Re: logical decoding : exceeded maxAllocatedDescs for .spill files  (Amit Khandekar <amitdkhan.pg@gmail.com>)
Responses Re: logical decoding : exceeded maxAllocatedDescs for .spill files  (Amit Khandekar <amitdkhan.pg@gmail.com>)
List pgsql-hackers
On Fri, Nov 22, 2019 at 7:38 PM Amit Khandekar <amitdkhan.pg@gmail.com> wrote:
>
> On Fri, 22 Nov 2019 at 4:26 PM, Amit Kapila <amit.kapila16@gmail.com> wrote:
>>
>> I think this is exactly the reason for the problem.  In my test [1],
>> the error "permission denied" occurred when I second time executed
>> pg_logical_slot_get_changes() which means on first execution the
>> unlink would have been successful but the files are still not removed
>> as they were not closed. Then on second execution, it gets an error
>> "Permission denied" when it again tries to unlink files via
>> ReorderBufferCleanupSerializedTXNs().
>>
>>
>> .
>> > But what you are seeing is "Permission denied" errors. Not sure why
>> > unlink() is failing.
>> >
>>
>> In your test program, if you try to unlink the file second time, you
>> should see the error "Permission denied".
>
>  I tested using the sample program and indeed I got the error 5 (access denied) when I called unlink the second
time.
>

So, what is the next step here?  How about if we somehow check whether
the file exists before doing unlink, say by using stat?  If that
doesn't work, I think we might need to go in the direction of tracking
file handles in some way, so that they can be closed during an abort.

-- 
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com



pgsql-hackers by date:

Previous
From: Amit Kapila
Date:
Subject: Re: [HACKERS] Block level parallel vacuum
Next
From: Michael Paquier
Date:
Subject: Re: Ought to use heap_multi_insert() for pg_attribute/dependinsertions?