| From: | Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> |
|---|---|
| To: | Dmitry Dolgov <9erthalion6(at)gmail(dot)com> |
| Cc: | Michael Paquier <michael(at)paquier(dot)xyz>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, James Coleman <jtc331(at)gmail(dot)com>, Pg Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
| Subject: | Re: remove spurious CREATE INDEX CONCURRENTLY wait |
| Date: | 2020-11-17 17:14:53 |
| Message-ID: | 20201117171453.GA19701@alvherre.pgsql |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
So I made the change to set the statusFlags with only LW_SHARED -- both
in existing uses (see 0002) and in the new CIC code (0003). I don't
think 0002 is going to have a tremendous performance impact, because it
changes operations that are very infrequent. But even so, it would be
weird to leave code around that uses exclusive lock when we're going to
add new code that uses shared lock for the same thing.
I still left the REINDEX CONCURRENTLY support in split out in 0004; I
intend to get the first three patches pushed now, and look into 0004
again later.
| Attachment | Content-Type | Size |
|---|---|---|
| v6-0001-indexcmds.c-reorder-function-prototypes.patch | text/x-diff | 1.6 KB |
| v6-0002-relax-lock-level-for-setting-statusFlags.patch | text/x-diff | 2.9 KB |
| v6-0003-CREATE-INDEX-CONCURRENTLY-don-t-wait-for-its-kin.patch | text/x-diff | 6.8 KB |
| v6-0004-Support-safe-flag-in-REINDEX-CONCURRENTLY.patch | text/x-diff | 4.1 KB |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Victor Yegorov | 2020-11-17 17:19:46 | Re: Deleting older versions in unique indexes to avoid page splits |
| Previous Message | Simon Riggs | 2020-11-17 17:14:14 | Re: Detecting File Damage & Inconsistencies |