Peerplays_Blockchain/peerplays_migrated

Fork 0

Resolve the issue when witness node is stopped with CTRL + C when some of the threads are not exit cleanly #394

New issue

Closed

opened 2022-06-18 21:10:51 +00:00 by hirunda · 12 comments

hirunda commented

2022-06-18 21:10:51 +00:00

(Migrated from gitlab.com)

Some of the fc::threads has the issue with race condition during close which is triggered by stopping the witness node with CTRL + C.

Explain where the issue is and find the workaround

Some of the fc::threads has the issue with race condition during close which is triggered by stopping the witness node with CTRL + C. Explain where the issue is and find the workaround

hirunda commented

2022-06-18 21:10:51 +00:00

(Migrated from gitlab.com)

assigned to @hirunda

hirunda commented

2022-07-04 11:40:41 +00:00

(Migrated from gitlab.com)

MR !134

hirunda commented

2022-07-04 11:42:50 +00:00

(Migrated from gitlab.com)

Instruction for testing after review pass:

Start sync with the mainnet
Interrupt sync with CTRL + C
Observe message at the end: Witness node is closed and turned off
Repeat the test ~50 times

Instruction for testing after review pass: 1) Start sync with the mainnet 2) Interrupt sync with CTRL + C 3) Observe message at the end: Witness node is closed and turned off 4) Repeat the test ~50 times

serkixenos commented

2022-07-04 17:53:46 +00:00

(Migrated from gitlab.com)

mentioned in issue #384

hirunda commented

2022-07-10 20:05:21 +00:00

(Migrated from gitlab.com)

assigned to @prandnum

prandnum commented

2022-07-12 18:17:39 +00:00

(Migrated from gitlab.com)

@hirunda Segmentation fault happened on the 36th ctrl+c operation. coredump, logs, witness binary, and blockchain databases are available here https://gitlab.com/PBSA/documentation/working-documents/-/tree/develop/394

CC: @serkixenos

prandnum commented

2022-08-01 16:30:50 +00:00

(Migrated from gitlab.com)

mentioned in issue #409

prandnum commented

2022-12-23 09:54:49 +00:00

(Migrated from gitlab.com)

@hirunda should we change the label back to pending?

vampik commented

2023-03-16 06:19:33 +00:00

(Migrated from gitlab.com)

This is currently unsolvable due to the complex threads which are fired for receiving blocks.

Davor added some check points on which we are exiting cleanly and most of the time we have clean exit.

But when you try for example 50 times, sometimes you will have an ugly print-out. Note that even in that situation, there is no database corruption.

This is currently unsolvable due to the complex threads which are fired for receiving blocks. Davor added some check points on which we are exiting cleanly and most of the time we have clean exit. But when you try for example 50 times, sometimes you will have an ugly print-out. Note that even in that situation, there is no database corruption.

vampik commented

2023-03-16 06:19:44 +00:00

(Migrated from gitlab.com)

Closing the issue.

vampik commented

2023-03-16 06:20:01 +00:00

(Migrated from gitlab.com)

assigned to @vampik

vampik (Migrated from gitlab.com) closed this issue

2023-03-16 06:20:02 +00:00

prandnum commented

2023-03-19 07:22:10 +00:00