Closed lacosteque closed 2 years ago
Hello, Please share complete unit.log.
Hi The error did not happen again, if it does I will give you the full log. But at the moment I left docker and just deployed it on the server, maybe the errors were because of docker. At the moment Unit works without errors and does not crash.
Hi
My log is flooded with the following messages
2022/02/16 22:53:45 [warn] 350#350 port #0 for pid 68427 remove failed
2022/02/16 22:53:45 [warn] 350#350 port #0 for pid 91709 remove failed
2022/02/16 22:53:45 [warn] 350#350 port #0 for pid 166042 remove failed
2022/02/16 22:53:45 [warn] 350#350 port #0 for pid 82073 remove failed
2022/02/16 22:53:45 [warn] 350#350 port #0 for pid 187702 remove failed
2022/02/16 22:53:45 [warn] 350#350 port #0 for pid 493825 remove failed
2022/02/16 22:53:45 [warn] 350#350 port #0 for pid 84002 remove failed
2022/02/16 22:53:45 [warn] 350#350 port #0 for pid 393899 remove failed
What is the problem? How do I fix it?
The unit itself works without errors.
Installed on a server paired with nginx without containers.
Warnings are one of the symptoms. When isolated PID number reach prototype process host PID, the prototype crashed.
The workaround - disable PID isolation. The patch - attached gh633_proto_isolation.patch.gz.
Removed from the configuration.
"Isolation." {
{ "namespaces": {
{ "cgroup": true,
{ "credential": true,
{ "mount": true,
{ "network": false }
{ "uname": true
}
},
now the log is full
2022/02/17 20:17:35 [info] 158027#158027 "newbon.ru" application started 294194 2022/02/17 20:17:51 [notice] 156992#156992 app process (isolated 158027) exited with code 0 2022/02/17 20:18:05 [info] 158063#158063 "newbon.ru" application started 2022/02/17 20:18:20 [notice] 156992#156992 app process (isolated 157957) exited with code 0 2022/02/17 20:18:43 [info] 158158#158158 "newbon.ru" application started 2022/02/17 20:18:54 [notice] 156990#156990 app process (isolated 158137) exited with code 0 2022/02/17 20:18:59 [notice] 156992#156992 app process (isolated 158063) exited with code 0 2022/02/17 20:19:11 [info] 158206#158206 "newbon.ru" application started 2022/02/17 20:19:26 [notice] 156992#156992 app process (isolated 158158) exited with code 0
Is it possible to install the patch, only when building the package or can it be installed on the installed application? Thank you!
It is required to rebuild main Unit binary (unitd
) to apply the patch.
It is required to rebuild main Unit binary (
unitd
) to apply the patch.
it's sad )) Can a unit be used with this error? The fact that the log grows, I will survive.
Versions:
The test environment is running wordpress with nginx & nginx unit. The load on the host, create only search bots. There is no other traffic.
But I noticed that Unit crashes after some time and Nginx gives 504 Gateway Time-out.
Here is an example of morning log at the moment when Unit crashed.
The only thing that helps is to completely restart the containers via podman-compose up
unit.conf
New unit.log
nginx.log
Perhaps you need to configure the config more correctly? And of course I do not particularly like such a number of warnings and alert, I would like to see, only, info messages in logs :)
Thank you!