| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> |
| Cc: | Amit Kapila <amit(dot)kapila16(at)gmail(dot)com>, Pavan Deolasee <pavan(dot)deolasee(at)gmail(dot)com>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Index corruption with CREATE INDEX CONCURRENTLY |
| Date: | 2017-02-06 17:30:29 |
| Message-ID: | 10113.1486402229@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> writes:
> Tom Lane wrote:
>> Better to fix the callers so that they don't have the assumption you
>> refer to. Or maybe we could adjust the API of RelationGetIndexAttrBitmap
>> so that it returns all the sets needed by a given calling module at
>> once, which would allow us to guarantee they're consistent.
> Note that my "interesting attrs" patch does away with these independent
> bitmaps (which was last posted by Pavan as part of his WARM set). I
> think we should fix just this bug now, and for the future look at that
> other approach.
BTW, if there is a risk of the assertion failure that Amit posits,
it seems like it should have happened in the tests that Pavan was doing
originally. I'd sort of like to see a demonstration that it can actually
happen before we spend any great amount of time fixing it.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Beena Emerson | 2017-02-06 17:39:50 | Re: increasing the default WAL segment size |
| Previous Message | Magnus Hagander | 2017-02-06 17:29:20 | Re: [PATCH] Rename pg_switch_xlog to pg_switch_wal |