[OpenIndiana-discuss] Access to ZFS viz CIFS from windows regularly hangs.

[Gordon Ross]

Have you done a snoop (or wireshark) capture to see what the box is
doing during the pause?
It's possible that it's trying to talk to an AD server...

On Tue, Jun 12, 2012 at 9:52 AM, John McEntee
<jmcentee at stirling-dynamics.com> wrote:
> I am having problems with a openindiana storage server I have built am I am
> trying to track down the cause to fix it. The current symptoms are seen from
> all windows clients (both 7 and XP) that will report an error stating.
>
>
>
> Path File is not accessible. The specified network name is no longer
> available.
>
>
>
> Another symptom  is windows explorer hangs and the user has to wait for it
> to some back.
>
>
>
> Just waiting a while ( a few minutes) and the box comes back.
>
>
>
> I  currently think the root cause is in openindiana somewhere but am at a
> bit of a loss. I have tried many things and have still not fixed it. I think
> the box is lightly loaded for the hardware spec but kernel load increases to
> 40% when a zfssnap is taking place.
>
>
>
> Hardware spec.
>
> 2 x Xeon E6520 cpus
>
> 48 GB RAM
>
> Intel HC5520 motherboard
>
> 3 x LSI SAS 9211-8i  cards
>
>
>
> Currently on openindiana 148
>
>
>
> The box is joined to a windows 2003 domain.
>
>
>
> Zpool tank is 3 way mirror of 7 x 3TB hitachi disk (using 21 disks in total,
> zpool size of 19 TB, ) with 2 x SSD   8GB ZIL  on each and 140GB L2ARC on
> each, default checksum, no dedup and no compression.
>
>
>
> Server operates as a windows home directory for 58 users (some laptops users
> so just a backup location), a main shared drive for the company of 120
> users.
>
> It is also a nfs server to a vmware vsphere 4 server hosting 10 virtual
> machines.
>
>
>
> There are only 8 active production file systems, and 12 backup file systems
> from other hosts (done out of hours).
>
>
>
> Zpool iostat peaks at about 35 MB for the pool mostly around the 0 to 7 MB
> level.
>
>
>
> Turning of time-sliderd does not stop the problem. (backups run out of
> hours)
>
>
>
> A  dtrace -n 'sched:::off-cpu { @[execname]=count()}'
>
> Used to give a sched count in the 6 to 7 fiqures over 3 seconds, but  turing
> apci off with
>
> #eeprom acpi-user-options=0x8
>
> Reduced this to 5 figures.
>
>
>
> What can I do to identify the problem to be able to fix it?
>
>
>
> Thanks
>
>
>
> John
>
>
>
> Other information:
>
>
>
> dtrace -n 'sched:::off-cpu { @[execname]=count()}'
>
> dtrace: description 'sched:::off-cpu ' matched 3 probes
>
> ^C
>
>
>
>  gconfd-2                                                          2
>
>  idmapd                                                            2
>
>  inetd                                                             2
>
>  nscd                                                              2
>
>  sendmail                                                          2
>
>  svc.startd                                                        2
>
>  gnome-power-mana                                                  3
>
>  fmd                                                               4
>
>  sshd                                                              4
>
>  devfsadm                                                          6
>
>  fsflush                                                           7
>
>  nfsmapid                                                          7
>
>  ntpd                                                              7
>
>  dtrace                                                           13
>
>  Xorg                                                             17
>
>  gdm-simple-greet                                                 17
>
>  svc.configd                                                      71
>
>  smbd                                                            113
>
>  time-sliderd                                                    138
>
>  zpool-rpool                                                     597
>
>  nfsd                                                            918
>
>  zpool-tank                                                     1968
>
>  sched                                                         80542
>
>
>
> # echo hz/D | sudo mdb -k
>
> hz:
>
> hz:             100
>
>
>
> # echo apic_timer::print apic_timer_t | sudo mdb -k
>
> {
>
>    mode = 0
>
>    apic_timer_enable_ops = oneshot_timer_enable
>
>    apic_timer_disable_ops = oneshot_timer_disable
>
>    apic_timer_reprogram_ops = oneshot_timer_reprogram
>
> }
>
>
>
> _______________________________________________________________________
>
> The contents of this e-mail and any attachment(s) are strictly confidential and are solely for the person(s) at the e-mail address(es) above. If you are not an addressee, you may not disclose, distribute, copy or use this e-mail, and we request that you send an e-mail to admin at stirling-dynamics.com and delete this e-mail.  Stirling Dynamics Ltd. accepts no legal liability for the contents of this e-mail including any errors, interception or interference, as internet communications are not secure.  Any views or opinions presented are solely those of the author and do not necessarily represent those of Stirling Dynamics Ltd. Registered In England No. 2092114 Registered Office: 26 Regent Street, Clifton, Bristol. BS8 4HG
> VAT no. GB 464 6551 29
> _______________________________________________________________________
>
> This e-mail has been scanned for all viruses MessageLabs.
> _______________________________________________
> OpenIndiana-discuss mailing list
> OpenIndiana-discuss at openindiana.org
> http://openindiana.org/mailman/listinfo/openindiana-discuss



-- 
Gordon Ross <gwr at nexenta.com>
Nexenta Systems, Inc.  www.nexenta.com
Enterprise class storage for everyone