rpc.lockd: errno = 70!

[Tapani Tarvainen]

Does anyone know what this error message means:

/usr/sbin/rpc.lockd: fcntl (local_lock) : errno = 70!

L1000, HP-UX 11.11, NFS server that is showing rather abysmal performance. No obvious resources are running low that I can see (I've got Glance+, but it doesn't show anything unusual).

Any ideas would be appreciated.

[Sundar_7]

I dont exactly know what is this

But why dont U try incrasing nflocks

Kernel Parameter

Sundar

[linuxfan]

Hi Tapani,

Not sure if it would help or not but found this doc for 10.X

For more information look at the document A5474915 in the knowledge base. I am pasting the document for your convenience

/Begin /
Problem Description

After I rebooted my NFS server, I began getting this
rpc.lockd logging error:

fcntl (local_lock) error = 70

What should I do?

Configuration Info

Operating System - HP-UX
Version - 10.20
Hardware System - HP9000
Series - K410

Solution

Use these steps:

1. kill rpc.statd rpc.lockd

2. rm -f /var/statmon/sm

3. rm -f /var/statmon/sm.bak

4. Run these two commands on both the server and client:

/usr/sbin/rpc.statd
/usr/sbin/rpc.lockd

Now the server will not keep trying to reestablish these mount
points, and everything will work fine.

/End/

-HTH
I am RU

[Tapani Tarvainen]

Some further thoughts: in /usr/include/sys/errno.h error 70 is "Stale NFS file handle".
If that is the case, it'd have to be some client - any idea how to track down which one?

- The problem isn't nflocks (plenty of those, according to Glance+ it never uses more than about 45% of what's available), and killing lockd & statd, removing /var/statmon/sm and restarting the daemons doesn't help. (It does help in another problem that keeps occurring in the same server every now and then, namely clients being unable to get file locks over NFS. I would also very much like to know why that happens.)

[Brian Hackley]

Tapani,

Without getting into a detailed support call, the procedure that I would follow would be to first ensure you have the NFS patch for 11i installed + dependencies. PHNE_23502. It fixes a large number of rpc.lockd issues.

Should that fail to resolve the issue, put rpc.lockd and rpc.statd into debug logging mode. Get their PIDS, use kill -17 PID and look in /var/adm/rpc.statd.log and rpc.lockd.log to ensure debug logging is on. Start a nettl trace nettl -tn all -e ns_ls_ip -tm 99999 -f /var/adm/crash/raw0 . Reproduce the problem. Turn off the nettl trace nettl -tf -e all, turn off debug logging for statd and lockd with another kill -17, verify it is turned off by look in the log file. Format the nettl trace after you look in the deub log files to determine which which client is of interest.
e.g.
put in filterfile IP of client of interest for example 192.1.1.1

filter ip_saddr 192.1.1.1
filter ip_daddr 192.1.1.1


netfmt -Nln1Tc filterfile -f /var/adm/crash/raw0.TRC0 > /var/adm/crash/fmt0-1

Hope all that helps,

-> Brian Hackley