Update kine for multiple fixes

[brandond]

Update kine for:

• fix: close mysql connection after create database kine#366
• Kine skips compact intervals if compaction fails to complete kine#357
• k3s database size grow due to slow compact process #10626 (reply in thread)
• k3s server create 100% CPU load #11251 (reply in thread)

I was able to reproduce database is locked errors after 10 minute of load testing v1.31.2-k3s1 using a small cluster with 1 kine+sqlite server and 2 agents, using https://github.com/brandond/jobloader:

NAME           STATUS   ROLES                  AGE   VERSION
k3s-agent-1    Ready    <none>                 13m   v1.31.2+k3s1
k3s-agent-2    Ready    <none>                 13m   v1.31.2+k3s1
k3s-server-1   Ready    control-plane,master   13m   v1.31.2+k3s1

INFO[0300] COMPACT compactRev=0 targetCompactRev=2 currentRev=17811
INFO[0300] COMPACT deleted 0 rows from 2 revisions in 1.482438ms - compacted to 2/17811

INFO[0600] COMPACT compactRev=2 targetCompactRev=1002 currentRev=38398
INFO[0600] COMPACT deleted 336 rows from 1000 revisions in 137.39489ms - compacted to 1002/38398
INFO[0600] COMPACT compactRev=1002 targetCompactRev=2002 currentRev=38403
ERRO[0600] Compact failed: failed to compact to revision 2002: database is locked

Another side effect of this is that the sqlite database file gets quite large after a while since compaction never succeeds and the WAL is never checkpointed:

root@k3s-server-1:~# ls -lah /var/lib/rancher/k3s/server/db/
total 6.2G
drwx------ 3 root root 4.0K Nov 11 22:50 .
drwx------ 8 root root 4.0K Nov 11 22:50 ..
drwx------ 2 root root 4.0K Nov 11 22:50 etcd
-rw-r--r-- 1 root root 1.4G Nov 12 00:42 state.db
-rw-r--r-- 1 root root 9.5M Nov 12 00:42 state.db-shm
-rw-r--r-- 1 root root 4.8G Nov 12 00:42 state.db-wal

[ShylajaDevadiga]

Validated the issue on k3s version v1.31.2+k3s-62caa4a8

On a 1 kine+sqlite server and 2 agents, using https://github.com/brandond/jobloader:

$ kubectl get nodes
NAME               STATUS   ROLES                  AGE     VERSION
ip-172-31-11-90    Ready    control-plane,master   3h39m   v1.30.2+k3s1
ip-172-31-14-168   Ready    <none>                 3h38m   v1.30.2+k3s1
ip-172-31-15-212   Ready    <none>                 3h37m   v1.30.2+k3s1

Results from reproducing the issue

ubuntu@ip-172-31-11-90:~$ k3s -v
k3s version v1.30.2+k3s1 (aa4794b3)
go version go1.22.4
ubuntu@ip-172-31-11-90:~$ journalctl -u k3s |grep locked
Nov 12 19:58:25 ip-172-31-11-90 k3s[1486]: time="2024-11-12T19:58:25Z" level=info msg="Bootstrap key locked for initial create"
Nov 12 20:47:59 ip-172-31-11-90 k3s[1486]: Trace[1872650200]: ---"watchCache locked acquired" 594ms (20:47:59.622)
Nov 12 20:48:16 ip-172-31-11-90 k3s[1486]: Trace[1490692442]: ---"watchCache locked acquired" 503ms (20:48:16.534)
Nov 12 20:53:30 ip-172-31-11-90 k3s[1486]: time="2024-11-12T20:53:30Z" level=error msg="Compact failed: failed to compact to revision 15102: database is locked"
Nov 12 21:03:26 ip-172-31-11-90 k3s[1486]: time="2024-11-12T21:03:26Z" level=error msg="Compact failed: failed to compact to revision 16102: database is locked"
Nov 12 21:13:26 ip-172-31-11-90 k3s[1486]: time="2024-11-12T21:13:26Z" level=error msg="Compact failed: failed to compact to revision 16102: database is locked"

Results from validation using latest commit on master

$ k3s -v
k3s version v1.31.2+k3s-62caa4a8 (62caa4a8)
go version go1.22.8
$ journalctl -u k3s |grep locked
Nov 12 23:10:27 ip-172-31-11-212 k3s[337358]: time="2024-11-12T23:10:27Z" level=info msg="Bootstrap key locked for initial create"
ubuntu@ip-172-31-11-212:~$