Percona Server moved to https://jira.percona.com/projects/PS

IMPORT TABLESPACE and undo tablespace truncate may get stuck indefinitely with a writing workload in parallel

Bug #1585095 reported by Laurynas Biveinis
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Percona Server moved to https://jira.percona.com/projects/PS
Status tracked in 5.7
5.5
Invalid
Undecided
Unassigned
5.6
Invalid
Undecided
Unassigned
5.7
Fix Released
High
Laurynas Biveinis
Percona Server moved to https://jira.percona.com/projects/PS 5.7.13-6

Bug Description

IMPORT TABLESPACE and undo tablespace truncate both call at some point

 log_make_checkpoint_at(LSN_MAX, TRUE);

which calls

 while (!log_preflush_pool_modified_pages(lsn)) {

with lsn == LSN_MAX, which, in PS with parallel doublewrite does not do any query-thread flushing itself, but just calls

 buf_flush_wait_flushed(new_oldest);

with new_oldest == LSN_MAX. To complete this wait with this target LSN value, buf_flush_wait_flushed will only go to the next buffer pool instance once the current one is clean. This wait becomes unbounded if there's a write workload in background which constantly dirties pages.

Laurynas Biveinis (laurynas-biveinis)
tags: added: regression xtradb
Revision history for this message
Laurynas Biveinis (laurynas-biveinis) wrote :

https://github.com/percona/percona-server/pull/541

Revision history for this message
Shahriyar Rzayev (rzayev-sehriyar) wrote :

Percona now uses JIRA for bug reports so this bug report is migrated to: https://jira.percona.com/browse/PS-985

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.
Subscribing...

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.