Problem with the Storage Event Logger (cmaeventd) PSP 8.31, RHEL 5.3 x86_64

[hmakripodis]

Hi to all,

We have problem with the Storage Event Logger (cmaeventd) daemon. After some days it 'eats' the one cpu core to 100% .

Details:

OS:                             Redhat Enterprise 5.3 x86_64

HP server model:    ProLiant BL460c G6
HP support pack installed: 8.31
Software Package:  hp-snmp-agents-8.3.0.27-24

I strace'd the cmaeventd process during the problematic operation and I get the below:

[pid 30658] access("/proc/driver/cciss/cciss0", F_OK) = 0
[pid 30658] open("/dev/cciss/c0d0", O_RDONLY) = 8
[pid 30658] ioctl(8, CCISS_PASSTHRU, 0xff8236c0) = -1 ENOMEM (Cannot allocate memory)
[pid 30658] ioctl(8, CCISS_PASSTHRU, 0xff8236c0) = -1 EINVAL (Invalid argument)
[pid 30658] close(8)                    = 0
[pid 30658] time(NULL)                  = 1275294056
[pid 30658] stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2245, ...}) = 0
[pid 30658] stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2245, ...}) = 0
[pid 30658] stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2245, ...}) = 0
[pid 30658] socket(PF_FILE, SOCK_DGRAM, 0) = 8
[pid 30658] fcntl64(8, F_SETFD, FD_CLOEXEC) = 0
[pid 30658] connect(8, {sa_family=AF_FILE, path="/dev/log"...}, 110) = 0
[pid 30658] send(8, "<13>May 31 11:20:56 cmaeventd[30"..., 39, MSG_NOSIGNAL) = 39
[pid 30658] close(8)                    = 0
[pid 30658] access("/proc/driver/cciss/cciss0", F_OK) = 0
[pid 30658] open("/dev/cciss/c0d0", O_RDONLY) = 8
[pid 30658] ioctl(8, CCISS_PASSTHRU, 0xff8236c0) = -1 ENOMEM (Cannot allocate memory)
[pid 30658] ioctl(8, CCISS_PASSTHRU, 0xff8236c0) = -1 EINVAL (Invalid argument)
[pid 30658] close(8)                    = 0
[pid 30658] time(NULL)                  = 1275294056
[pid 30658] stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2245, ...}) = 0
[pid 30658] stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2245, ...}) = 0
[pid 30658] stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2245, ...}) = 0
[pid 30658] socket(PF_FILE, SOCK_DGRAM, 0) = 8
[pid 30658] fcntl64(8, F_SETFD, FD_CLOEXEC) = 0
[pid 30658] connect(8, {sa_family=AF_FILE, path="/dev/log"...}, 110) = 0
[pid 30658] send(8, "<13>May 31 11:20:56 cmaeventd[30"..., 39, MSG_NOSIGNAL) = 39
[pid 30658] close(8)                    = 0
[pid 30658] access("/proc/driver/cciss/cciss0", F_OK) = 0
[pid 30658] open("/dev/cciss/c0d0", O_RDONLY) = 8
[pid 30658] ioctl(8, CCISS_PASSTHRU, 0xff8236c0) = -1 ENOMEM (Cannot allocate memory)
[pid 30658] ioctl(8, CCISS_PASSTHRU, 0xff8236c0) = -1 EINVAL (Invalid argument)
[pid 30658] close(8)                    = 0
[pid 30822] <... restart_syscall resumed> ) = 0
[pid 30822] rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
[pid 30822] rt_sigaction(SIGCHLD, NULL, {SIG_DFL, [], SA_STACK|SA_NODEFER|SA_RESETHAND|0xc30280}, 8) = 0
[pid 30822] rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
[pid 30822] nanosleep({1, 0},  <unfinished ...>
[pid 30658] time(NULL)                  = 1275294056
[pid 30658] stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2245, ...}) = 0
[pid 30658] stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2245, ...}) = 0
[pid 30658] stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=2245, ...}) = 0
[pid 30658] socket(PF_FILE, SOCK_DGRAM, 0) = 8
[pid 30658] fcntl64(8, F_SETFD, FD_CLOEXEC) = 0
[pid 30658] connect(8, {sa_family=AF_FILE, path="/dev/log"...}, 110) = 0
[pid 30658] send(8, "<13>May 31 11:20:56 cmaeventd[30"..., 39, MSG_NOSIGNAL) = 39
[pid 30658] close(8)                    = 0
[pid 30658] access("/proc/driver/cciss/cciss0", F_OK) = 0
[pid 30658] open("/dev/cciss/c0d0", O_RDONLY) = 8
[pid 30658] ioctl(8, CCISS_PASSTHRU, 0xff8236c0) = -1 ENOMEM (Cannot allocate memory)
[pid 30658] ioctl(8, CCISS_PASSTHRU, 0xff8236c0) = -1 EINVAL (Invalid argument)

The problem is temporary solved by restarting the service 'hp-snmp-agents' and re-appears again in a few days later..

Any advice ?

Thank you