RPC: Timed out

[bhawani123]

Hi Guys,

I have 10 number of ultra 60 and ultra 10 systems and each system has solaris 2.7 operating system.
My one system /usr/local file system is shared and it is mounted on each system through /etc/auto_direct file.Some time I get the following error message in client system but same time another system use the same file system.
error message----

automountd[198]: ultrasparc:/usr/local server not responding: RPC: Timed out
Apr 2 21:30:01 dcmtaj last message repeated 2 times
Apr 3 21:29:22 dcmtaj automountd[198]: ultrasparc:/usr/local server not responding: RPC: Timed out
Apr 3 21:29:42 dcmtaj last message repeated 1 time

Can anybody help me out?why i am getting the error message.If some NFS tunning is required please tell me the way to do this.

Regd,
Ajay

 

[Igaduma]

Hi there,

You can check nfs stats with

nfsstat -a

Look for badcalls & the like,
check with rpcinfo what ports are used,

use nfsstat -r to check rpc info.
The percentages returned should be self-explanatory to determine whats the problem

Also, /etc/dfs/sharetab is used for the types of sharing you will allow on what systems/directories, I put the whole shebang on anon, works ok.

I had the same thing occuring very randomly,
turned out it was the router, which was configured to run at 100hdx while the nic's where hardcoded at 100fdx.

I

 

[mikeclark]

Could be collisions on the network causing the servers to drop to half duplex which can run like a dog. The best thing is to lock your network card at 100mb FDX and your switch port to 100mb FDX as well.
Some server and switch combinations have problems with autonegotiation, so locking them to 100mv FDX stops this.

 

[bhawani123]

Thanks Igaduma,

I am attaching output of nfsstat -a and nfsstat -r command.
Please tell me how to resolve this.

ultrasparc# nfsstat -a|more

Server nfs_acl:
Version 2: (17424939 calls)
null getacl setacl getattr access
0 0% 66 0% 0 0% 17137533 98% 287340 1%
Version 3: (131379 calls)
null getacl setacl
0 0% 131379 100% 0 0%

Client nfs_acl:
Version 2: (1 calls)
null getacl setacl getattr access
0 0% 0 0% 0 0% 1 100% 0 0%
Version 3: (66623 calls)
null getacl setacl
0 0% 66287 99% 336 0%


ultrasparc# nfsstat -r|more

Server rpc:
Connection oriented:
calls badcalls nullrecv badlen xdrcall
301305212 0 0 0 0
dupchecks dupreqs
10799335 0
Connectionless:
calls badcalls nullrecv badlen xdrcall
100020 5 0 0 5
dupchecks dupreqs
0 0

Client rpc:
Connection oriented:
calls badcalls badxids timeouts newcreds
11005263 795 426 102 0
badverfs timers cantconn nomem interrupts
0 0 267 0 426
Connectionless:
calls badcalls retrans badxids timeouts
164184 14 0 0 13

Regd,
Ajay Lohani

 

[Igaduma]

Hi there,
I think u should go back and check the network settings or the nfs deamon number, it normally only runs max 16 connections, if i'm not mistaken.
You can change that number in inet, if you need more, run more.

As mikeclark says, could be network collisions, since timeouts are occuring and such. Most of the time thats the problem, with autoneg routers (cisco) the speed goes up & down, gotta hardcode the port setting, both client & router.

out-of-the-box testing with 1 server and 1 client over a correctly set-up network hardly ever results in errors,