Re: nfs4 client hangs
On 08/22/2012 02:07 AM, Barbara Krasovec wrote: Hello! Nfs4 client randomly hangs on a file or directory (for instance when performing df or find) and after a few seconds continues. When preforming strace, it hangs on getdents system call. Errors on the server side: Aug 19 21:29:08 host kernel: nfs: server host2 not responding, still trying Aug 19 21:31:17 host kernel: nfs: server host2 OK After a while the load on client and server side increases intensively, then system on the client side crashes. Reboot required. We use SL6.3, the same problem occurred on the following kernels: 2.6.32-220.17.1.el6.x86_64 2.6.32-279.2.1.el6.x86_64 and also on: 3.5.2-1.el6.elrepo.x86_64 3.3.2-1.el6.elrepo.x86_64 The only difference is that when we are using 3.x kernel from elrepo, the system crash never appears, just client hangs. Mount options: Flags: rw,relatime,vers=4.0,rsize=1048576,wsize=1048576,namlen=255,hard,noresvport,proto=tcp,port=0,timeo=600,retrans=2,sec=sys,clientaddr=IP,local_lock=none,addr=IP Nfs utils version: nfs-utils-1.2.3-15.el6_2.1.x86_64 Any ideas? Thanks, Barbara This doesn't look quite the same, but we had problems with hung tasks since updating to 6.3 on a couple machines. Filed a bug here: https://bugzilla.redhat.com/show_bug.cgi?id=851673 -- Orion Poplawski Technical Manager 303-415-9701 x222 NWRA, Boulder Office FAX: 303-415-9702 3380 Mitchell Lane or...@nwra.com Boulder, CO 80301 http://www.nwra.com
Re: nfs4 client hangs
On 08/23/2012 01:44 PM, Klaus Steinberger wrote: Hi, Nfs4 client randomly hangs on a file or directory (for instance when performing df or find) and after a few seconds continues. When preforming strace, it hangs on getdents system call. Errors on the server side: Aug 19 21:29:08 host kernel: nfs: server host2 not responding, still trying Aug 19 21:31:17 host kernel: nfs: server host2 OK After a while the load on client and server side increases intensively, then system on the client side crashes. Reboot required. Mount options: Flags: rw,relatime,vers=4.0,rsize=1048576,wsize=1048576,namlen=255,hard,noresvport,proto=tcp,port=0,timeo=600,retrans=2,sec=sys,clientaddr=IP,local_lock=none,addr=IP SL5 or SL6 on Server side? We had many hungs with SL5 as server but with kerberos auth, never tried NFS4 with sec=sys. Sincerly, Klaus We have sl6 as a server and sl6 and sl5 clients... Nfs hangs occur on sl6 client... (at least I didn't notice any on sl5 clients)... Soft mount isn't an option because of the nature of the service running on the machine. Maybe I could try automount if there will be any difference.. Thanks, Barbara
nfs4 client hangs
Hello! Nfs4 client randomly hangs on a file or directory (for instance when performing df or find) and after a few seconds continues. When preforming strace, it hangs on getdents system call. Errors on the server side: Aug 19 21:29:08 host kernel: nfs: server host2 not responding, still trying Aug 19 21:31:17 host kernel: nfs: server host2 OK After a while the load on client and server side increases intensively, then system on the client side crashes. Reboot required. We use SL6.3, the same problem occurred on the following kernels: 2.6.32-220.17.1.el6.x86_64 2.6.32-279.2.1.el6.x86_64 and also on: 3.5.2-1.el6.elrepo.x86_64 3.3.2-1.el6.elrepo.x86_64 The only difference is that when we are using 3.x kernel from elrepo, the system crash never appears, just client hangs. Mount options: Flags: rw,relatime,vers=4.0,rsize=1048576,wsize=1048576,namlen=255,hard,noresvport,proto=tcp,port=0,timeo=600,retrans=2,sec=sys,clientaddr=IP,local_lock=none,addr=IP Nfs utils version: nfs-utils-1.2.3-15.el6_2.1.x86_64 Any ideas? Thanks, Barbara
Re: nfs4 client hangs
On 08/22/2012 05:07 PM, Barbara Krasovec wrote: Nfs4 client randomly hangs on a file or directory (for instance when performing df or find) and after a few seconds continues. When preforming strace, it hangs on getdents system call. Errors on the server side: Aug 19 21:29:08 host kernel: nfs: server host2 not responding, still trying Aug 19 21:31:17 host kernel: nfs: server host2 OK After a while the load on client and server side increases intensively, then system on the client side crashes. Reboot required. Mount options: Flags: rw,relatime,vers=4.0,rsize=1048576,wsize=1048576,namlen=255,hard,noresvport,proto=tcp,port=0,timeo=600,retrans=2,sec=sys,clientaddr=IP,local_lock=none,addr=IP I haven't had this problem with several systems mounting /home, but sometime before 2.6.32-220 we switched to using automount instead of fstab. Have you tried playing with the options to see if one of them is poisonous (hard, retrans, local_lock, etc.)? Do you have the same problem using automount?