kernel26-2.6.30.5-1 breaks nfs-server

jcci · 2009-08-26 13:33:03

This is really weird. On one nfs-server the update does no harm, but on another the nfs-server fails after update. Wehen trying a restart of all services rpcbind and nfs-common are OK, but nfs-server shows "fail" at rpc.nfsd. Re-installing the kernel didn't help, but downgrading to kernel26-2.6.30.4-1 solved the problem.
I don't know where the difference between the 2 machines is. /etc/exports is same in formar, /etc/conf.d/nfs-common.conf and /etc/conf.d/nfs-server.conf are default in all cases.

shining · 2009-08-26 14:13:45

you cannot get any addition information about the error? "fail" does not say anything.
if you run /usr/sbin/rpc.nfsd directly, it fails too? no error message? nothing in the log or in dmesg?

jcci · 2009-08-27 01:21:55

This is what I found in errors.log:
Aug 26 14:49:53 123-Server rpcbind: cannot bind * on udp6: Invalid argument
Aug 26 14:49:54 123-Server nfsd[1922]: nfssvc: writing fds to kernel failed: errno 97 (Address family not supported by protocol)
Aug 26 14:49:54 123-Server nfsd[1922]: nfssvc: writing fds to kernel failed: errno 97 (Address family not supported by protocol)
Aug 26 14:49:54 123-Server nfsd[1922]: nfssvc: Address already in use

...I just don't know what it means!

shining · 2009-08-27 06:29:38

Are you using ipv6? If not maybe you could try to disable it?
http://wiki.archlinux.org/index.php/IPv … the_Module

jcci · 2009-08-28 03:47:52

Thank you for the suggestion. Unfortunately it doesn't help.
On an Arch client the following message is displayed when trying a mount:
"mount.nfs: an incorrect mount option was specified"
No matter there was ANY option specified. An Ubuntu client claims an unregistered service.
The problem is absolutely reproducable and dissapers when downgrading to kernel26-2.6.30.4-1.

AndyRTR · 2009-08-28 04:29:25

There were no changes in our kernel config files.

Make sure your whole system is up to date as well as the mirror you are using. Also check for related .pacnew config files you may have missed to merge. Otherwise look out for an upstream regression at the LKML.

http://www.kernel.org/pub/linux/kernel/ … g-2.6.30.5 - there's at least one nfs related change you may locally revert for testing.

My private nfs server is still working well.

jcci · 2009-08-28 05:48:57

My other NFS server did work well until just now. All NFS-services crashed on the second machine while everybody were working on it and there was no way of solving the problem. In this case a downgrade didn't help at all so this issue might be not related just to the kernel. It was the solution for the first server which failed!
Fortunately I always keep a proofed-to-be-stable snapshot of the server just on another HDD. So restarting the server with the older Arch system (2.6.30-1) and everything works.
It should be impossible to be a singular case as my 2 server failed independent of each other, with different pattern, but at about the same time of upgrade.

AndyRTR · 2009-08-28 06:40:40

you might think about using the kernel26-lts on your servers after some testing

Last edited by AndyRTR (2009-08-28 06:41:14)

jcci · 2009-08-28 08:14:11

Thanks for this idea! But ext4 support should be stable.
Anyway, so far I don't know how to approach this problem.
Just some idea: Is it possible that Samba interferes with RPC? I found lots of entries in "errors.log" which look not good.

Aug 28 09:04:57 FS smbd[2618]: [2009/08/28 09:04:57, 0] lib/util_sock.c:get_peer_addr_internal(1676)
Aug 28 09:04:57 FS smbd[2618]: getpeername failed. Error was Transport endpoint is not connected
Aug 28 09:04:57 FS smbd[2618]: [2009/08/28 09:04:57, 0] lib/util_sock.c:write_data(1136)
Aug 28 09:04:57 FS smbd[2618]: [2009/08/28 09:04:57, 0] lib/util_sock.c:get_peer_addr_internal(1676)
Aug 28 09:04:57 FS smbd[2618]: getpeername failed. Error was Transport endpoint is not connected
Aug 28 09:04:57 FS smbd[2618]: write_data: write failure in writing to client 0.0.0.0. Error Connection reset by peer
Aug 28 09:04:57 FS smbd[2618]: [2009/08/28 09:04:57, 0] smbd/process.c:srv_send_smb(74)
Aug 28 09:04:57 FS smbd[2618]: Error writing 4 bytes to client. -1. (Transport endpoint is not connected)
Aug 28 09:59:50 FS smbd[2738]: [2009/08/28 09:59:50, 0] lib/util_sock.c:write_data(1136)
Aug 28 09:59:50 FS smbd[2738]: [2009/08/28 09:59:50, 0] lib/util_sock.c:get_peer_addr_internal(1676)
Aug 28 09:59:50 FS smbd[2738]: getpeername failed. Error was Transport endpoint is not connected
Aug 28 09:59:50 FS smbd[2738]: write_data: write failure in writing to client 0.0.0.0. Error Connection reset by peer
Aug 28 09:59:50 FS smbd[2738]: [2009/08/28 09:59:50, 0] smbd/process.c:srv_send_smb(74)
Aug 28 09:59:50 FS smbd[2738]: Error writing 4 bytes to client. -1. (Transport endpoint is not connected)
Aug 28 10:26:01 FS smbd[2639]: [2009/08/28 10:26:01, 0] lib/util_sock.c:read_socket_with_timeout(939)
Aug 28 10:26:01 FS smbd[2639]: [2009/08/28 10:26:01, 0] lib/util_sock.c:get_peer_addr_internal(1676)
Aug 28 10:26:01 FS smbd[2639]: getpeername failed. Error was Transport endpoint is not connected
Aug 28 10:26:01 FS smbd[2639]: read_socket_with_timeout: client 0.0.0.0 read error = No route to host.
Aug 28 11:04:21 FS kernel: ata4.00: exception Emask 0x0 SAct 0x1 SErr 0x0 action 0x6 frozen
Aug 28 11:04:21 FS kernel: ata4.00: cmd 61/08:00:30:da:de/00:00:07:00:00/40 tag 0 ncq 4096 out
Aug 28 11:04:21 FS kernel: res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
Aug 28 11:04:21 FS kernel: ata4.00: status: { DRDY }
Aug 28 11:22:35 FS smbd[2470]: [2009/08/28 11:22:35, 0] smbd/service.c:make_connection(1288)
Aug 28 11:22:35 FS smbd[2470]: max_yu (::ffff:192.168.111.28) couldn't find service maxy
Aug 28 11:31:18 FS smbd[2470]: [2009/08/28 11:31:18, 0] smbd/nttrans.c:call_nt_transact_ioctl(1989)
Aug 28 11:31:18 FS smbd[2470]: call_nt_transact_ioctl(0x900bc): Currently not implemented.
Aug 28 11:34:51 FS smbd[2470]: [2009/08/28 11:34:51, 0] smbd/service.c:make_connection(1288)
Aug 28 11:34:51 FS smbd[2470]: max_yu (::ffff:192.168.111.28) couldn't find service maxy
Aug 28 11:50:03 FS smbd[2470]: [2009/08/28 11:50:03, 0] smbd/service.c:make_connection(1288)
Aug 28 11:50:03 FS smbd[2470]: max_yu (::ffff:192.168.111.28) couldn't find service maxy
Aug 28 12:05:14 FS smbd[2470]: [2009/08/28 12:05:14, 0] smbd/service.c:make_connection(1288)
Aug 28 12:05:14 FS smbd[2470]: max_yu (::ffff:192.168.111.28) couldn't find service maxy
Aug 28 12:30:24 FS smbd[2470]: [2009/08/28 12:30:24, 0] smbd/service.c:make_connection(1288)
Aug 28 12:30:24 FS smbd[2470]: max_yu (::ffff:192.168.111.28) couldn't find service maxy

Both server show lots of Samba entries in "error.log". In this case I could be impossibly the only one with this problem.

Arch Linux

#1 2009-08-26 13:33:03

kernel26-2.6.30.5-1 breaks nfs-server

#2 2009-08-26 14:13:45

Re: kernel26-2.6.30.5-1 breaks nfs-server

#3 2009-08-27 01:21:55

Re: kernel26-2.6.30.5-1 breaks nfs-server

#4 2009-08-27 06:29:38

Re: kernel26-2.6.30.5-1 breaks nfs-server

#5 2009-08-28 03:47:52

Re: kernel26-2.6.30.5-1 breaks nfs-server

#6 2009-08-28 04:29:25

Re: kernel26-2.6.30.5-1 breaks nfs-server

#7 2009-08-28 05:48:57

Re: kernel26-2.6.30.5-1 breaks nfs-server

#8 2009-08-28 06:40:40

Re: kernel26-2.6.30.5-1 breaks nfs-server

#9 2009-08-28 08:14:11

Re: kernel26-2.6.30.5-1 breaks nfs-server

Board footer