JFTR: Upgraded to 8.1.3P2 yesterday, everything good again – dedupes are no longer causing performance issues, so this was the solution.
Bye,
Alexander Griesser
System-Administrator
ANEXIA Internetdienstleistungs GmbH
Telefon: +43-5-0556-320
Telefax: +43-5-0556-500
E-Mail:
ag@anexia.at
Web:
http://www.anexia.at
Anschrift Hauptsitz Klagenfurt: Feldkirchnerstraße 140, 9020 Klagenfurt
Geschäftsführer: Alexander Windbichler
Firmenbuch: FN 289918a | Gerichtsstand: Klagenfurt | UID-Nummer: AT U63216601
Von: toasters-bounces@teaparty.net [mailto:toasters-bounces@teaparty.net]
Im Auftrag von Aaron Lewis
Gesendet: Samstag, 19. Oktober 2013 02:22
An: toasters@teaparty.net
Betreff: RE: AW: Dedup killing performance (again)
I’ll have to agree as well, we’re recommending all of our customers go to either 8.1.3 or 8.2P3. We’ve had several people suffer from the dedupe
fingerprinting issue in 8.1.2 and now there’s a serious data corruption bug (#723354) in 8.2 pre-P3.
Bug description:
o burt
723354 “Illegal container vbn panic while loading a BLKR MOVED buf”; SW, dense, sev-2. A problem in all current 8.2 releases (8.2RC1, 8.2GA, 8.2P1,
8.2P2, and any D-patch derivatives) which may result in WAFL inconsistencies on systems running either 7-Mode or clustered Data ONTAP if they have dedupe-enabled volumes and there is repeated overwriting of 4KB blocks while a deduplication process is running
against that volume.
A is included in Data ONTAP 8.2P3 and existing customers using dedupe on 8.2 are advised to disable dedupe now to avoid further exposure, and then upgrade to 8.2P3 as soon as is operationally feasable (at which time dedupe can be re-enabled). It is also recommended
that customers using dedupe on a pre-8.2 release of Data ONTAP that are considering upgrading to Data ONTAP 8.2 upgrade to 8.2P3 (or later) to avoid this issue.
From:
toasters-bounces@teaparty.net [mailto:toasters-bounces@teaparty.net]
On Behalf Of Alexander Griesser
Sent: Friday, October 18, 2013 5:13 PM
To: Fred Grieco; Mark Allen
Cc: Peter-Paul Witta; toasters@teaparty.net
Subject: AW: AW: Dedup killing performance (again)
Alright, I’m almost convinced to go for 8.1.3 now – will read through the changelogs tomorrow morning; right now, everything is fine again, fingerprint recovery
has been finished a few minutes after my last mail regarding this issue and I’m now deleting some old snapshots to regain that space which should be fine for now.
I’ve disabled any future sis schedules for now until I know how I want to proceed with that situation.
Thanks for all your fast help as always
J
Bye,
Alexander Griesser
System-Administrator
ANEXIA Internetdienstleistungs GmbH
Telefon: +43-5-0556-320
Telefax: +43-5-0556-500
E-Mail:
ag@anexia.at
Web:
http://www.anexia.at
Anschrift Hauptsitz Klagenfurt: Feldkirchnerstraße 140, 9020 Klagenfurt
Geschäftsführer: Alexander Windbichler
Firmenbuch: FN 289918a | Gerichtsstand: Klagenfurt | UID-Nummer: AT U63216601
Von: Fred Grieco [mailto:fredgrieco@yahoo.com]
Gesendet: Samstag, 19. Oktober 2013 02:11
An: Alexander Griesser; Mark Allen
Cc: Peter-Paul Witta; toasters@teaparty.net
Betreff: Re: AW: Dedup killing performance (again)
|
I have 8.1.3p1 on a 3140 and a 3270 (7 mode) with no dedupe performance issues.
Sent from Yahoo! Mail on Android |
From:
Mark Allen <mmmjka@gmail.com>;
To: Alexander Griesser <ag@anexia.at>;
Cc: Peter-Paul Witta <paul.witta@cubit.at>;
<toasters@teaparty.net>;
Subject: Re: AW: Dedup killing performance (again)
Sent: Sat, Oct 19, 2013 12:00:06 AM