mirror of
https://github.com/fluencelabs/redis
synced 2025-03-20 09:30:55 +00:00
Panic when we are sending an error to our master/slave.
Related to #5135, see discussion there.
This commit is contained in:
parent
e03358c0d9
commit
afc7e08a20
@ -349,6 +349,11 @@ void addReplyErrorLength(client *c, const char *s, size_t len) {
|
|||||||
serverLog(LL_WARNING,"== CRITICAL == This %s is sending an error "
|
serverLog(LL_WARNING,"== CRITICAL == This %s is sending an error "
|
||||||
"to its %s: '%s' after processing the command "
|
"to its %s: '%s' after processing the command "
|
||||||
"'%s'", from, to, s, cmdname);
|
"'%s'", from, to, s, cmdname);
|
||||||
|
/* Here we want to panic because when an instance is sending an
|
||||||
|
* error to another instance in the context of replication, this can
|
||||||
|
* only create some kind of offset or data desynchronization. Better
|
||||||
|
* to catch it ASAP and crash instead of continuing. */
|
||||||
|
serverPanic("Continuing is unsafe: replication protocol violation.");
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
|
Loading…
x
Reference in New Issue
Block a user