zfs discuss - Jun 2011 - SE11 express Encryption on - > errors in the pool after Scrub

Hi,

I am testing Solaris Express 11 with napp-it on two machines. In both 
cases the same problem: Enabling encryption on a folder, filling it with 
data will result in errors indicated by a subsequent scrub. I did not 
find the topic on the web, but also not experiences shared by people 
using encryption on SE11 express. Advice would be highly appreciated.

Best
Otto

> From: zfs-discuss-bounces at opensolaris.org [mailto:zfs-discuss-
> bounces at opensolaris.org] On Behalf Of Thomas Hobbes
> 
> I am testing Solaris Express 11 with napp-it on two machines. In both
> cases the same problem: Enabling encryption on a folder, filling it with
> data will result in errors indicated by a subsequent scrub. I did not
> find the topic on the web, but also not experiences shared by people
> using encryption on SE11 express. Advice would be highly appreciated.
Most likely you''ll have the same experience with/without encryption. 
Look
for a hardware problem.

If you don''t find a hardware problem, look deeper.  For example, maybe
there''s only a few bad sectors on disk, and you happened to hit them
when
you ran encryption.  So if you don''t find the problem repeating the
same
test without encryption ...  Use a larger data set, up to the size of the
entire pool.

Probably the fastest and easiest...  Run something like memetest86 to test
all your RAM.  Are you using ECC ram?  You could probably skip the memtest
safely if you have ECC.

If that still doesn''t yield any results,  try stressing &
overheating the
CPU doing unrelated work at the same time while you''re repeating your
disk
test unencrypted.  16-threaded infinite random number generator dumping to
the null bucket, while you''re running your disk test without
encryption.

That should be a pretty exhaustive process to check your CPU, memory,
motherboard & buses, disk controllers, and disks.  I bet you''ll
find some
hardware issue.

Dear Edward,

I will run a couple of tests (you suggested) over night, tomorrow and send  
all in one.

Concerning hardware:
Fileserver "Micro": HP Microserver, 8 GB ECC RAM, , Intel Dual NIC
Card,
Raidsonic Bay ST1000-2-S2C (instead of DVD, HDD with OS in that)
4x1TB non 4k WD Green HDDs

Esxi-Server "Romaschka": Intel Xeon UP L3406 (30 Watt), 16 GB Kingston
ECC
non reg., Supermicro X8SIE-LN4F. Esxi and Solaris VM on WD Raptor 300 GB.  
5x WD Green EARS 2TB on Intel SASUC8I per passtrough to Solaris VM  
(napp-it al in one concept).

No clue if I said but I am Political Scientist by profession - so not a  
nerd and quite new to Solaris (some people however wonder why i have a two  
meter 19 inch rack in the office...). So might be that sometimes I have to  
ask a second time.

Cheers





Am 05.06.2011, 17:03 Uhr, schrieb Edward Ned Harvey  
<opensolarisisdeadlongliveopensolaris at nedharvey.com>:
>> From: zfs-discuss-bounces at opensolaris.org [mailto:zfs-discuss-
>> bounces at opensolaris.org] On Behalf Of Thomas Hobbes
>>
>> I am testing Solaris Express 11 with napp-it on two machines. In both
>> cases the same problem: Enabling encryption on a folder, filling it
with
>> data will result in errors indicated by a subsequent scrub. I did not
>> find the topic on the web, but also not experiences shared by people
>> using encryption on SE11 express. Advice would be highly appreciated.
>
> Most likely you''ll have the same experience with/without
encryption.
> Look
> for a hardware problem.
>
> If you don''t find a hardware problem, look deeper.  For example,
maybe
> there''s only a few bad sectors on disk, and you happened to hit
them when
> you ran encryption.  So if you don''t find the problem repeating
the same
> test without encryption ...  Use a larger data set, up to the size of the
> entire pool.
>
> Probably the fastest and easiest...  Run something like memetest86 to  
> test
> all your RAM.  Are you using ECC ram?  You could probably skip the  
> memtest
> safely if you have ECC.
>
> If that still doesn''t yield any results,  try stressing &
overheating the
> CPU doing unrelated work at the same time while you''re repeating
your
> disk
> test unencrypted.  16-threaded infinite random number generator dumping  
> to
> the null bucket, while you''re running your disk test without
encryption.
>
> That should be a pretty exhaustive process to check your CPU, memory,
> motherboard & buses, disk controllers, and disks.  I bet
you''ll find some
> hardware issue.
>

-- 
Erstellt mit Operas revolution?rem E-Mail-Modul: http://www.opera.com/mail/

On 06/04/11 13:52, Thomas Hobbes wrote:
> I am testing Solaris Express 11 with napp-it on two machines. In both
> cases the same problem: Enabling encryption on a folder, filling it with
> data will result in errors indicated by a subsequent scrub. I did not
> find the topic on the web, but also not experiences shared by people
> using encryption on SE11 express. Advice would be highly appreciated.
If you are doing the scrub when the encryption keys are not present it 
is possible you are hitting a known (and very recently fixed in the 
Solaris 11 development gates) bug.

If you have an operating systems support contract with Oracle you should 
be able to log a support ticket and request a backport of the fix for CR 
6989185.

-- 
Darren J Moffat

Ok,

I tested it. It made two Scrubs with open encrypted folders. No issues  
anymore. Thanks for the hint. Hope that will be fixed for all soon.

Cheers


Am 06.06.2011, 11:54 Uhr, schrieb Darren J Moffat  
<darrenm at opensolaris.org>:
> On 06/04/11 13:52, Thomas Hobbes wrote:
>> I am testing Solaris Express 11 with napp-it on two machines. In both
>> cases the same problem: Enabling encryption on a folder, filling it
with
>> data will result in errors indicated by a subsequent scrub. I did not
>> find the topic on the web, but also not experiences shared by people
>> using encryption on SE11 express. Advice would be highly appreciated.
>
> If you are doing the scrub when the encryption keys are not present it  
> is possible you are hitting a known (and very recently fixed in the  
> Solaris 11 development gates) bug.
>
> If you have an operating systems support contract with Oracle you should  
> be able to log a support ticket and request a backport of the fix for CR  
> 6989185.
>

-- 
Erstellt mit Operas revolution?rem E-Mail-Modul: http://www.opera.com/mail/