possible error with priority node connection about monero HOT 15 CLOSED

This is identical to #8132 . I think my PR #7345 could fix the issue, but I have not heard back from the reporter in the other thread (and I have not seen this on my local node). Listing it as a priority node should be a red herring, but cannot say for certain until the problem is found.

from monero.

aight, running #7345 . i'll check the logs in a day to see if they are full of those stacktrace

from monero.

looking good so far

Wednesday?
user@user-ProLiant-DL325-Gen10:~/.bitmonero$ grep stacktrace bitmonero.log* | wc -l
6106

Thu Nov 16 15:54:05 EST 2023
user@user-ProLiant-DL325-Gen10:~/.bitmonero$ grep stacktrace bitmonero.log* | wc -l
5143

Fri Nov 17 06:08:20 EST 2023
user@user-ProLiant-DL325-Gen10:~/.bitmonero$ grep stacktrace bitmonero.log* | wc -l
3767

Fri Nov 17 17:24:03 EST 2023
user@user-ProLiant-DL325-Gen10:~/.bitmonero$ grep stacktrace bitmonero.log* | wc -l
2376

from monero.

Sat Nov 18 06:57:30 EST 2023
user@user-ProLiant-DL325-Gen10:~/.bitmonero$ grep stacktrace bitmonero.log* | wc -l
1272

Sun Nov 19 08:10:57 EST 2023
user@user-ProLiant-DL325-Gen10:~/.bitmonero$ grep stacktrace bitmonero.log* | wc -l
202

Running the same command on the xmrchain box:
xmrchain@Ubuntu-1804-bionic-64-minimal:~/.bitmonero$ grep stacktrace bitmonero.log* | wc -l
213102

uptime for my local box
uptime 3d 12h 27m 53s

I'll give this a week for stability, and then test it on xmrchain

from monero.

Of course i realize now i should be grepping for weak_ptr

xmrchain@Ubuntu-1804-bionic-64-minimal:~/.bitmonero$ grep weak_ptr bitmonero.log* | wc -l
9055

user@user-ProLiant-DL325-Gen10:~/.bitmonero$ grep weak_ptr bitmonero.log* | wc -l
9

and the ones on proliant are all from 11/16, which was before I added this PR.

from monero.

@Gingeropolous so does this appear to help? Because the grep is still returning results (old logs?).

from monero.

yeah, its old logs. It looks like its helping the situation.

yep, currently there are 0 instances

Mon Nov 20 21:32:12 EST 2023
user@user-ProLiant-DL325-Gen10:~/.bitmonero$ grep weak_ptr bitmonero.log* | wc -l
0

from monero.

on homebox,
uptime 7d 9h 44m 18s

patch considered stable on homebox

Now testing on xmrchain

Prior to pulling in 7345:

All logs

Thu Nov 23 11:30:53 AM CET 2023
xmrchain@Ubuntu-1804-bionic-64-minimal:~/.bitmonero$ grep weak_ptr bitmonero.log* | wc -l
12749

Recent Logs

xmrchain@Ubuntu-1804-bionic-64-minimal:~/.bitmonero$ grep weak_ptr bitmonero.log | wc -l
2162

from monero.

Recent Logs

xmrchain@Ubuntu-1804-bionic-64-minimal:~/.bitmonero$ grep weak_ptr bitmonero.log | wc -l
2162

So it's still happening? Unfortunately I am not seeing this on my local box. Does it require a higher log level than default?

from monero.

Recent Logs

xmrchain@Ubuntu-1804-bionic-64-minimal:~/.bitmonero$ grep weak_ptr bitmonero.log | wc -l
2162

So it's still happening? Unfortunately I am not seeing this on my local box. Does it require a higher log level than default?

Answered my own question (forgot the log prints level)

from monero.

sorry im making this confusing. Those were the recent logs before the patch were applied.

as far as I can tell, its not happening. The recent log (bitmonero.log) shows there are no new instances, and the total log (bitmonero.log* , which includes all the logrotated logs) is showing a decreasing number compared to before the patch was applied

Height: 3024783/3024783 (100.0%) on mainnet, not mining, net hash 2.41 GH/s, v16, 127(out)+113(in) connections, uptime 0d 16h 10m 34s
xmrchain@Ubuntu-1804-bionic-64-minimal:~/.bitmonero$ date
Fri Nov 24 03:42:24 AM CET 2023
xmrchain@Ubuntu-1804-bionic-64-minimal:~/.bitmonero$ grep weak_ptr bitmonero.log | wc -l
0
xmrchain@Ubuntu-1804-bionic-64-minimal:~/.bitmonero$ grep weak_ptr bitmonero.log* | wc -l
6792

The patch seems successful

from monero.

and yeah, I run log-level 1

from monero.

Ok, sounds like this patch fixes the issue. The problem is probably a race-condition to the destructor of the connection object. I'm not aware of another fix, besides using a weak_ptr to stop the abuse of shared_ptr (which is what my PR does).

from monero.

fixed by #7345

from monero.

Now I just have to get reviewers ;)

from monero.