Messages in this thread | | | Date | Mon, 25 Jan 1999 11:25:18 +0100 | From | Lenart Gabor <> | Subject | 2.2.0-final NFS problems ... |
| |
(VM of 2.2.0-final is GREAT, but this mail is not about this ...)
We've got a server running Linux 2.2.0-final and acting as NFS server for many other Linux machines. Some hours ago knfsd put an 'oppps' message onto terminal. It must be said that nobody could mount the server before the opps either ... Kernel modules nfsd,lockd,sunrpc stuck in memory and we can't rmmod them after stopping knfs server via stopping its rc script. Then I tried to start user space NFS server, but it didn't worked to. On server side I strace-ed rpc.mountd :
select(1024, [4 5], NULL, NULL, NULL) = ? ERESTARTNOHAND (To be restarted) --- SIGALRM (Alarm clock) --- time([917258751]) = 917258751 alarm(5) = 0 sigreturn() = ? (mask now []) select(1024, [4 5], NULL, NULL, NULL) = ? ERESTARTNOHAND (To be restarted) --- SIGALRM (Alarm clock) --- time([917258756]) = 917258756 alarm(5) = 0 sigreturn() = ? (mask now []) select(1024, [4 5], NULL, NULL, NULL) = 1 (in [5]) accept(5, {sin_family=AF_INET, sin_port=htons(670), sin_addr=inet_addr("193.6.41.138")}, [16]) = 0 select(1024, [0 4 5], NULL, NULL, NULL) = 1 (in [0]) select(1024, [0 4 5], NULL, NULL, {35, 0}) = 1 (in [0], left {35, 0}) read(0, "\200\0\0T5\2576\254\0\0\0\0\0\0\0"..., 4000) = 88 time(NULL) = 917258756 lstat("/usr", {st_mode=0, st_size=0, ...}) = 0 time([917258756]) = 917258756 getpid() = 1602 sigaction(SIGPIPE, {0x400b37e0, [], 0}, {SIG_IGN}) = 0 send(3, "<29>Jan 25 11:05:56 mountd[1602]"..., 81, 0) = 81 sigaction(SIGPIPE, {SIG_IGN}, NULL) = 0 stat("/usr", {st_mode=0, st_size=0, ...}) = 0 lstat("/", {st_mode=S_ISUID|01, st_size=0, ...}) = 0 lstat("/usr", {st_mode=0, st_size=0, ...}) = 0 lstat("/", {st_mode=S_ISUID|01, st_size=0, ...}) = 0 open("/", O_RDONLY|O_NONBLOCK) = 1 fcntl(1, F_SETFD, FD_CLOEXEC) = 0 fstat(1, {st_mode=S_ISUID|01, st_size=0, ...}) = 0 lseek(1, 0, SEEK_CUR) = 0 getdents(1, /* 19 entries */, 3933) = 312 close(1) = 0 time([917258756]) = 917258756 open("/etc/hosts", O_RDONLY) = 1 fcntl(1, F_GETFD) = 0 fcntl(1, F_SETFD, FD_CLOEXEC) = 0 fstat(1, {st_mode=0, st_size=0, ...}) = 0 mmap(0, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x4010b000 read(1, "127.0.0.1\tlocalhost\n193.6.41.1"..., 4096) = 1125 close(1) = 0 munmap(0x4010b000, 4096) = 0 stat("/etc/rmtab", {st_mode=0, st_size=0, ...}) = 0 time([917258756]) = 917258756 getpid() = 1602 sigaction(SIGPIPE, {0x400b37e0, [], SA_STACK|0x62aa0}, {SIG_IGN}) = 0 send(3, "<29>Jan 25 11:05:56 mountd[1602]"..., 73, 0) = 73 sigaction(SIGPIPE, {SIG_IGN}, NULL) = 0 write(0, "\200\0\0<5\2576\254\0\0\0\1\0\0\0"..., 64) = 64 select(1024, [0 4 5], NULL, NULL, NULL) = 1 (in [0]) select(1024, [0 4 5], NULL, NULL, {35, 0}) = 1 (in [0], left {35, 0}) read(0, "", 4000) = 0 close(0) = 0 select(1024, [4 5], NULL, NULL, NULL) = ? ERESTARTNOHAND (To be restarted) --- SIGALRM (Alarm clock) --- time([917258761]) = 917258761 alarm(5) = 0 sigreturn() = ? (mask now []) select(1024, [4 5], NULL, NULL, NULL) = ? ERESTARTNOHAND (To be restarted)
[... many times ...]
What is ERESTARTNOHAND ? man page of select() does not contain any information on this error.
It can be important that we haven't got any problem until a power fault occured and we can't reach our name server, but any hosts eg in /etc/exports are given by theit IP addresses so I don't understand the problem ...
It's sad, but we haven't got the opps message because of the broken remote syslog mechanism casued by non-reachable server we're logging to ...
root@hal2000:~# dpkg -S `which rpc.mountd` nfs-server: /usr/sbin/rpc.mountd root@hal2000:~# dpkg -l nfs-server Desired=Unknown/Install/Remove/Purge | Status=Not/Installed/Config-files/Unpacked/Failed-config/Half-installed |/ Err?=(none)/Hold/Reinst-required/X=both-problems (Status,Err: uppercase=bad) ||/ Name Version Description +++-===============-==============-============================================ ii nfs-server 2.2beta37-1 User space NFS server.
(In /usr/src/linux/Changes 2.2beta40 is mentioned but I haven't had any problem using 2.2.0-preX kernels with this nfs-server package on my Linux for months)
----
Another question about NFS : If I remove some files or directories on an exported directory at the server side, client which try to access produces MANY messages into syslog which is very ugly. I think it would be better to simply ignore these requests.
---[ LGB/DC ]------------[ University Of Veszprém ]------[ Lénárt Gábor ]--- "The truth is out there" "We're l{ea,i}ving together" "The future is dark." ---[ 88/422022(4602) ]---[ http://lgb.hal.vein.hu ]---------[ 87/477074 ]---
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |