Bug 178231 - [nfs] 8.3 nfsv4 client reports "nfsv4 client/server protocol prob err=10026"
Summary: [nfs] 8.3 nfsv4 client reports "nfsv4 client/server protocol prob err=10026"
Status: Open
Alias: None
Product: Base System
Classification: Unclassified
Component: kern (show other bugs)
Version: Unspecified
Hardware: Any Any
: Normal Affects Only Me
Assignee: freebsd-bugs (Nobody)
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2013-04-29 06:40 UTC by Alex Strange
Modified: 2020-11-21 15:56 UTC (History)
1 user (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Alex Strange 2013-04-29 06:40:01 UTC
I have a busy web server which mounts two other systems over NFSv4 and writes files to them. All three systems are running FreeBSD 8.3 amd64.

dmesg frequently fills up with this:
nfsv4 client/server protocol prob err=10026
nfsv4 client/server protocol prob err=10026


which causes file writes to fail on specific worker processes. Killing the right process while the logging is going fixes the issue for a while, but this would cause requests to error so we'd prefer to avoid it.

nfs activity:
% nfsstat -w 1 -e
        GtAttr Lookup Rdlink   Read  Write Rename Access  Rddir

Client:   1106    880      0      0    424     60    880      0
Server:      0      0      0      0      0      0      0      0

nfs related sysctls:
vfs.newnfs.downdelayinitial: 12
vfs.newnfs.downdelayinterval: 30
vfs.newnfs.keytab_enctype: 1
vfs.newnfs.skip_wcc_data_onerr: 1
vfs.newnfs.nfs3_jukebox_delay: 10
vfs.newnfs.reconnects: 0
vfs.newnfs.bufpackets: 4
vfs.newnfs.callback_addr: 
vfs.newnfs.nfs4acl_enable: 1
vfs.newnfs.realign_count: 0
vfs.newnfs.realign_test: 0
vfs.newnfs.nfs_directio_allow_mmap: 1
vfs.newnfs.nfs_directio_enable: 0
vfs.newnfs.clean_pages_on_close: 1
vfs.newnfs.commit_on_close: 0
vfs.newnfs.prime_access_cache: 0
vfs.newnfs.access_cache_timeout: 60
vfs.newnfs.diskless_rootpath: 
vfs.newnfs.diskless_valid: 0
vfs.newnfs.nfs_ip_paranoia: 1
vfs.newnfs.defect: 0
vfs.newnfs.iodmax: 4
vfs.newnfs.iodmin: 0
vfs.newnfs.iodmaxidle: 120
vfs.newnfs.disable_checkutf8: 0
vfs.newnfs.server_max_nfsvers: 4
vfs.newnfs.server_min_nfsvers: 2
vfs.newnfs.nfs_privport: 0
vfs.newnfs.enable_locallocks: 0
vfs.newnfs.issue_delegations: 0
vfs.newnfs.commit_miss: 0
vfs.newnfs.commit_blks: 0
vfs.newnfs.mirrormnt: 1
vfs.newnfs.minthreads: 1
vfs.newnfs.maxthreads: 4
vfs.newnfs.threads: 0
vfs.newnfs.request_space_used: 0
vfs.newnfs.request_space_used_highest: 0
vfs.newnfs.request_space_high: 47185920
vfs.newnfs.request_space_low: 31457280
vfs.newnfs.request_space_throttled: 0
vfs.newnfs.request_space_throttle_count: 0

Applicable fstab:
boards:/boards /www/boards nfs rw,noatime,nfsv4,async,nolockd,late,intr,bg,timeout=5 0 0
images:/images /www/images nfs rw,noatime,nfsv4,async,nolockd,late,intr,bg,timeout=5 0 0
images:/thumbs /www/thumbs nfs rw,noatime,nfsv4,async,nolockd,late,intr,bg,timeout=5 0 0

How-To-Repeat: Wait.
Comment 1 Mark Linimon freebsd_committer freebsd_triage 2013-05-20 04:29:21 UTC
Responsible Changed
From-To: freebsd-bugs->freebsd-fs

Over to maintainer(s).
Comment 2 Rick Macklem freebsd_committer 2013-05-20 22:06:30 UTC
State Changed
From-To: open->feedback


I have emailed the reporter and asked him to test without 
the "intr" mount option, which could cause this. The error 
indicates that the client got "out of sync" with the server 
while doing open/lock operations, which are serialized via 
a sequence number.
Comment 3 Eitan Adler freebsd_committer freebsd_triage 2017-12-31 07:59:45 UTC
For bugs matching the following criteria:

Status: In Progress Changed: (is less than) 2014-06-01

Reset to default assignee and clear in-progress tags.

Mail being skipped
Comment 4 ml 2020-11-21 15:56:54 UTC
Hello.

I've got a 12.2 client connecting to a 12.2 server and I'm seeing this message a few times a day.
I'm not sure it leads to consequences, though.

I also rarely see the same message, but with error 10020 and "nfsv4 expired locks lost".

I don't use the intr option.



I'd be glad to help investigate this, but I need guidance.