On Tue, May 18, 2021 at 1:29 PM osumi.takami...@fujitsu.com <osumi.takami...@fujitsu.com> wrote: > > On Monday, May 17, 2021 6:45 PM Amit Kapila <amit.kapil...@gmail.com> wrote: > > > > We allow taking locks on system catalogs, so why prohibit > > user_catalog_tables? However, I agree that if we want plugins to acquire the > > lock on user_catalog_tables then we should either prohibit decoding of such > > relations or do something else to avoid deadlock hazards. > OK. > > Although we have not concluded the range of logical decoding of > user_catalog_table > (like we should exclude TRUNCATE command only or all operations on that type > of table), > I'm worried that disallowing the logical decoding of user_catalog_table > produces > the deadlock still. It's because disabling it by itself does not affect the > lock taken by TRUNCATE command. What I have in mind is an example below. > > (1) plugin (e.g. pgoutput) is designed to take a lock on user_catalog_table. > (2) logical replication is set up in synchronous mode. > (3) TRUNCATE command takes an access exclusive lock on the user_catalog_table. > (4) This time, we don't do anything for the TRUNCATE decoding. > (5) the plugin tries to take a lock on the truncated table > but, it can't due to the lock by TRUNCATE command. >
If you skip decoding of truncate then we won't invoke plugin API so step 5 will be skipped. > I was not sure that the place where the plugin takes the lock is in > truncate_cb > or somewhere else not directly related to decoding of the user_catalog_table > itself, > so I might be wrong. However, in this case, > the solution would be not disabling the decoding of user_catalog_table > but prohibiting TRUNCATE command on user_catalog_table in synchronous_mode. > If this is true, I need to extend an output plugin and simulate the deadlock > first > and remove it by fixing the TRUNCATE side. Thoughts ? > I suggest not spending too much time reproducing this because it is quite clear that it will lead to deadlock if the plugin acquires lock on user_catalog_table and we allow decoding of truncate. But if you want to see how that happens you can try as well. -- With Regards, Amit Kapila.
