I had problems, that some services appeared unsupported for a moment, but returned on next reading. This was especially bad for speed/delta oriented data. Looking at the agent logs I think I can see what's causing the problem:
013385:20060724:125201 Connection from [x]. Allowed servers [x]
013385:20060724:125201 Before read()
013385:20060724:125201 After read() 2 [1043]
013385:20060724:125201 Got line:vserver.in[dns]
io[disk_wio]
vfs.fs.size[/,used]
agent.ping
vserver.out[somacica]
system.swap.size[,free]
vfs.fs.size[/,used]
vserver.in[savik]
system.swap.size[,free]
net.if.in[eth0]
net.if.in[eth0]
vm.memory.size[buffers]
io[disk_io]
vserver.in[zitest]
vserver.out[dns]
io[disk_io]
system.users.num
vm.memory.size[buffers]
vserver.out[savik]
system.users.num
net.if.out[eth0]
net.if.out[eth0]
vm.memory.size[cached]
io[disk_rblk]
vserver.out[zitest]
vserver.in[dns]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
system.users.num
system.swap.size[,free]
system[procrunning]
agent.ping
system.users.num
vm.memory.size[shared]
proc.num[]
system[procrunning]
agent.ping
system.cpu.load[,avg5]
vm.memory.size[shared]
proc.num[]
system.swap.size[,free]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
io[disk_rblk]
vm.memory.size[buffers]
vm.memory.size[shared]
proc.num[]
system[procrunning]
agent.ping
system.cpu.load[,avg5]
vm.memory.size[shared]
proc.num[]
system.swap.size[,free]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
io[disk_rblk]
vm.memory.size[buffers]
vm.memory.size[shared]
vserver.in[zitest]
system[procrunning]
vserver.out[savserver.out[savik]
The last line is garbled (and I think that is the one that will turn out to be unsupported in that round of reads). This behaviour continues to the next read as it apparently starts with the end of the previous one:
013384:20060724:125231 Before read()
013384:20060724:125231 After read() 2 [1047]
013384:20060724:125231 Got line:vik]
system.cpu.load[,avg5]
net.tcp.service[ssh]
io[disk_wio]
io[disk_rio]
...
This seems like a s(n)printf problem, the garbling is exactly at the 1024th byte. I'm using v1.1.1 of both agent and server.
013385:20060724:125201 Connection from [x]. Allowed servers [x]
013385:20060724:125201 Before read()
013385:20060724:125201 After read() 2 [1043]
013385:20060724:125201 Got line:vserver.in[dns]
io[disk_wio]
vfs.fs.size[/,used]
agent.ping
vserver.out[somacica]
system.swap.size[,free]
vfs.fs.size[/,used]
vserver.in[savik]
system.swap.size[,free]
net.if.in[eth0]
net.if.in[eth0]
vm.memory.size[buffers]
io[disk_io]
vserver.in[zitest]
vserver.out[dns]
io[disk_io]
system.users.num
vm.memory.size[buffers]
vserver.out[savik]
system.users.num
net.if.out[eth0]
net.if.out[eth0]
vm.memory.size[cached]
io[disk_rblk]
vserver.out[zitest]
vserver.in[dns]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
system.users.num
system.swap.size[,free]
system[procrunning]
agent.ping
system.users.num
vm.memory.size[shared]
proc.num[]
system[procrunning]
agent.ping
system.cpu.load[,avg5]
vm.memory.size[shared]
proc.num[]
system.swap.size[,free]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
io[disk_rblk]
vm.memory.size[buffers]
vm.memory.size[shared]
proc.num[]
system[procrunning]
agent.ping
system.cpu.load[,avg5]
vm.memory.size[shared]
proc.num[]
system.swap.size[,free]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
system.cpu.load[,avg5]
io[disk_rblk]
vm.memory.size[buffers]
vm.memory.size[shared]
vserver.in[zitest]
system[procrunning]
vserver.out[savserver.out[savik]
The last line is garbled (and I think that is the one that will turn out to be unsupported in that round of reads). This behaviour continues to the next read as it apparently starts with the end of the previous one:
013384:20060724:125231 Before read()
013384:20060724:125231 After read() 2 [1047]
013384:20060724:125231 Got line:vik]
system.cpu.load[,avg5]
net.tcp.service[ssh]
io[disk_wio]
io[disk_rio]
...
This seems like a s(n)printf problem, the garbling is exactly at the 1024th byte. I'm using v1.1.1 of both agent and server.
Comment