My failure rate went up spectacularly overnight ... is it possible that some
changes checked in yesterday have broken things?
For example ...
$ pmlc -Dall
...
pmlc> show loggers @vm21
stmt=278, state=0, control=0, hostspec=vm21, pid=-2, port=-2
auxconnect.c:__pmGetAddrInfo(bozo) -> (null)
auxconnect.c:__pmGetAddrInfo(vm21) -> (null)
auxconnect.c:__pmHostEntFree(hostent=0x1f2c0b0) name=(nil) ((null))
addresses=0x1f2c330
auxconnect.c:__pmHostEntFree(hostent=0x1f2d0a0) name=(nil) ((null))
addresses=0x1f2e4f0
auxconnect.c:__pmGetAddrInfo(bozo) -> (null)
auxconnect.c:__pmGetAddrInfo(vm21) -> (null)
auxconnect.c:__pmHostEntFree(hostent=0x1f2d0a0) name=(nil) ((null))
addresses=0x1f2e760
auxconnect.c:__pmHostEntFree(hostent=0x1f2c0b0) name=(nil) ((null))
addresses=0x1f2e590
Error finding pmloggers on vm21: Unknown code gvAS 96
pmlc>
What's this "Unknown code gvAS 96" stuff?
On another system (code from a few days ago), the same operation produces ...
pmlc> show loggers @vm21
stmt=278, state=0, control=0, hostspec=vm21, pid=-2, port=-2
auxconnect.c:__pmGetAddrInfo(vm01) -> (null)
auxconnect.c:__pmGetAddrInfo(vm21) -> (null)
auxconnect.c:__pmHostEntFree(hostent=0x835d068) name=(nil) ((null))
addresses=0x835d160
auxconnect.c:__pmHostEntFree(hostent=0x835d1d0) name=(nil) ((null))
addresses=0x835e4d8
__pmConnectPMCD: trying __pmAuxConnectPMCDPort(vm21, 44321) ...
auxconnect.c:__pmGetAddrInfo(vm21) -> (null)
__pmSetSocketIPC: fd=3
IPC table fd(PDU version):
__pmDataIPC: fd=3, data=0x835ec10(sz=16)
__pmDataIPC: fd=3, data=0x835ec10(sz=16)
__pmDataIPC: fd=3, data=0x835ec10(sz=16)
auxconnect.c:__pmHostEntFree(hostent=0x835ebb0) name=(nil) ((null))
addresses=0x835d938
... (lots of tracing) ...
The following pmloggers are running on vm21:
primary (16147) 21909
Does this match any place changes have been made recently? My first guess
would be the Coverity-driven commits, but that is only a guess.
|