You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We're using ProxySQL version 1.4.14-percona-1.1 (on Debian) and today we encountered a crash on one of our ProxySQL hosts. Upon further investigation it revealed that the crash happened due to the watchdog missing 10 heartbeats on one of our clusters.
2020-01-13 11:12:24 MySQL_Monitor.cpp:1437:monitor_ping(): [ERROR] Server <some.borked.host>:3306 missed 3 heartbeats, shunning it and killing all the connections. Disabling other checks until the node comes back online.
2020-01-13 11:12:24 main.cpp:1091:main(): [ERROR] Watchdog: 4 threads missed a heartbeat
2020-01-13 11:12:27 main.cpp:1091:main(): [ERROR] Watchdog: 4 threads missed a heartbeat
2020-01-13 11:12:30 main.cpp:1091:main(): [ERROR] Watchdog: 4 threads missed a heartbeat
2020-01-13 11:12:33 main.cpp:1091:main(): [ERROR] Watchdog: 4 threads missed a heartbeat
2020-01-13 11:12:36 main.cpp:1091:main(): [ERROR] Watchdog: 4 threads missed a heartbeat
2020-01-13 11:12:39 main.cpp:1091:main(): [ERROR] Watchdog: 4 threads missed a heartbeat
2020-01-13 11:12:42 main.cpp:1091:main(): [ERROR] Watchdog: 4 threads missed a heartbeat
2020-01-13 11:12:45 main.cpp:1091:main(): [ERROR] Watchdog: 4 threads missed a heartbeat
2020-01-13 11:12:48 main.cpp:1091:main(): [ERROR] Watchdog: 4 threads missed a heartbeat
2020-01-13 11:12:51 main.cpp:1091:main(): [ERROR] Watchdog: 4 threads missed a heartbeat
2020-01-13 11:12:51 main.cpp:1095:main(): [ERROR] Watchdog: reached 10 missed heartbeats. Aborting!
proxysql: main.cpp:1096: int main(int, const char**): Assertion '0' failed.
Error: signal 6:
proxysql(_Z13crash_handleri+0x1a)[0x560b95850eda]
/lib/x86_64-linux-gnu/libc.so.6(+0x33060)[0x7f825ae30060]
/lib/x86_64-linux-gnu/libc.so.6(gsignal+0xcf)[0x7f825ae2ffff]
/lib/x86_64-linux-gnu/libc.so.6(abort+0x16a)[0x7f825ae3142a]
/lib/x86_64-linux-gnu/libc.so.6(+0x2be67)[0x7f825ae28e67]
/lib/x86_64-linux-gnu/libc.so.6(+0x2bf12)[0x7f825ae28f12]
proxysql(main+0x749)[0x560b9584ece9]
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf1)[0x7f825ae1d2e1]
proxysql(_start+0x2a)[0x560b9584f03a]
I find the issue by Ben Mildren (1685) the most interesting one as it's exactly what we encounter with one slight difference: it seems that the line of shunningthe host in question prior the watchdog failing is the cause of all four threads locking up.
We have a core dump of proxysql that we can look into with you and we can also share our configuration with you.
The text was updated successfully, but these errors were encountered:
Yes, sure, we can have a look at core dump, but currently our backlog on community bugs is quite busy, as paying customers get priority.
Speaking of priority, the issue reported by Ben (#1685) is fixed in a hotfix for 1.4.10 (on commit e5226f9), and it is fixed in 2.0.5 (bd9c88e). It never made into 1.4.x .
Note that the hotfix is dated May 2019, while the latest 1.4.x (1.4.15) was released in Feb 2019 .
In short, my first suggestion is to upgrade, or to apply the hotfix to 1.4.14 (we can do that for you).
We're using ProxySQL version 1.4.14-percona-1.1 (on Debian) and today we encountered a crash on one of our ProxySQL hosts. Upon further investigation it revealed that the crash happened due to the watchdog missing 10 heartbeats on one of our clusters.
I see we're not the only ones struggling with it:
#1685
#1701
#2103
#2217
#2286
#2361
I find the issue by Ben Mildren (1685) the most interesting one as it's exactly what we encounter with one slight difference: it seems that the line of shunningthe host in question prior the watchdog failing is the cause of all four threads locking up.
We have a core dump of proxysql that we can look into with you and we can also share our configuration with you.
The text was updated successfully, but these errors were encountered: