If you're getting RPC timeouts on writes, it is most likely because you have your NFS client mounting the filer via a soft mount. This is not recommended for this exact purpose. If there are network delays (and 98 times out of 100 they are network-based delays, in my experience), a soft mount just gives up after a timeout period. The default, a hard mount, will continue to retry until the op succeeds.
Tracing mountd is probably not going to help your solve your problem as mountd problems tend to be access/permission issues.
Now just changing the mount option from soft to hard may not (and probably will not) solve the underlying problem. You may end up exchanging RPC timeouts for NFS server not responding messages on your client. In that case you will still have to investigate why the response is taking too long. Typically on the filer side this is done via a packet trace (pktt) where you (or someone at NTAP if you open a case) will examine the filer-side packet trace and determine things like: Are the requests getting to the filer? And if so, how quickly is the filer responding to the requsts? Answering these questions will determine if you have a filer problem or a network-based problem.
-- Adam Fox adamfox@netapp.com
-----Original Message----- From: Zeitschel, Philipp [mailto:philipp.zeitschel@siemens.com] Sent: Friday, August 24, 2007 8:21 AM To: toasters@mathworks.com Subject: AW: nfs logging / debugging
Hi,
in the syslog messages in FilerView is nothing about nfs. in the /etc/log/ems is a littlebit about nfs, but only lines like this:
<LR d="20Aug2007 07:33:08" n="sodn02x" pn="sodn01x" t="1187587988" id="1184669953/106342" p="7" s="C=1U" o="ExportPool00" vf="erpvfiler"> <kern_syslog_msg_3 msg="Client erlm5p1x (xid 313175200), matches to erlm5p1x for misc access" vfilerName="erpvfiler"/> </LR> <LR d="20Aug2007 07:33:08" n="sodn02x" pn="sodn01x" t="1187587988" id="1184669953/106343" p="7" s="C=1U" o="ExportPool00" vf="erpvfiler"> <kern_syslog_msg_3 msg="Key erlm5p1x (xid 313175200), group ptdsiscon is not found in netgroup med for misc access" vfilerName="erpvfiler"/> </LR> <LR d="23Aug2007 17:14:28" n="sodn02x" pn="sodn01x" t="1187882068" id="1184669953/131313" p="7" s="C=1U" o="MNTPool08" vf=""> <kern_syslog_msg_3 msg="Client 146.254.168.188 (xid 102334387) in mount, does not have access rights to path /vol/vol0/delme_schaar" vfilerName=""/> </LR>
but this is not really usefull vor debugging. Is there no logfile, about the transactions?
here a little example:
the sunlogfile show me:
Aug 22 19:20:46 rd2c03 nfs: [ID 277861 kern.notice] NFS write error on host erpvfiler: error 145.
now i look at the /etc/log/ems:
there is no mathing entry to this timestamp +-5min
and one look in /etc/messages:
no entry from 19:00 to 20:00
can we say, if the filer had an error, there where entrys? can i say to the sunteam, it's not the filers fault?
greetings
Philipp
-----Ursprüngliche Nachricht----- Von: Holland, William L [mailto:HollandWL@state.gov] Gesendet: Freitag, 24. August 2007 13:03 An: Zeitschel, Philipp Betreff: RE: nfs logging / debugging
Look at the syslog messages in FilerView. They will look very similar in format to your sun messages.
-----Original Message----- From: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Zeitschel, Philipp Sent: Friday, August 24, 2007 5:46 AM To: toasters@mathworks.com Subject: AW: nfs logging / debugging
Hi,
i've got problems, with a sun machine, here are the log:
Aug 22 19:20:46 rd2c03 nfs: [ID 277861 kern.notice] NFS write error on host erpvfiler: error 145. Aug 22 19:20:46 rd2c03 nfs: [ID 702911 kern.notice] (file handle: 22018100 bcf22707 20080000 17bf48b 62c93907 7a7133b9 40000000 70a27100) Aug 22 19:20:46 rd2c03 nfs: [ID 664466 kern.notice] NFS write failed for server erpvfiler: error 5 (RPC: Timed out) Aug 22 19:20:46 rd2c03 last message repeated 2 times Aug 22 19:20:46 rd2c03 nfs: [ID 277861 kern.notice] NFS write error on host erpvfiler: error 145. Aug 22 19:20:46 rd2c03 last message repeated 2 times Aug 22 19:20:46 rd2c03 nfs: [ID 702911 kern.notice] (file handle: 22018100 bcf22707 20080000 17bf48b 62c93907 7a7133b9 40000000 70a27100) Aug 22 19:20:46 rd2c03 last message repeated 2 times Aug 22 19:20:46 rd2c03 nfs: [ID 664466 kern.notice] NFS write failed for server erpvfiler: error 5 (RPC: Timed out) Aug 22 19:20:46 rd2c03 last message repeated 1 time Aug 22 19:20:46 rd2c03 nfs: [ID 277861 kern.notice] NFS write error on host erpvfiler: error 145. Aug 22 19:20:46 rd2c03 nfs: [ID 664466 kern.notice] NFS write failed for server erpvfiler: error 5 (RPC: Timed out) Aug 22 19:20:46 rd2c03 nfs: [ID 277861 kern.notice] NFS write error on host erpvfiler: error 145. Aug 22 19:20:46 rd2c03 nfs: [ID 702911 kern.notice] (file handle: fc8f7200 c7e2b006 20080000 1e0617a 8573af12 7a7133b9 40000000 70a27100)
and so on....
i wanna see, what the filer says on this times, is it a network problem, or an nfs server problem. But where is the advanced (if there is one) nfs logging?
-----Ursprüngliche Nachricht----- Von: Blake Golliher [mailto:thelastman@gmail.com] Gesendet: Freitag, 24. August 2007 11:00 An: Zeitschel, Philipp Cc: toasters@mathworks.com Betreff: Re: nfs logging / debugging
It might help to know what problem you are having. Usually /etc/messages has enough information to get started on with most problems.
-Blake
On 8/24/07, Zeitschel, Philipp
philipp.zeitschel@siemens.com wrote:
Hi,
I'm new to this list, and my english is not very well...
i've got some nfs problems, so i wan't to read some
log's. With the
option nfs.mountd.trace on the filer logs some stuff to
/etc/log/ems,
is that right?
the ems log, looks like some pseudo xml, is there a
special tool to
analyze the file?
are there other possibilitys to debug nfs on the filer?
NetApp Release 7.2.3
greeting
Philipp