Hello everybody I have a Fas3204 HA in two seperate chassis , dataontap version 8.02P3 7mode. Many NFS servers are connected to it via iscsi connections. NFS servers are VMware mashines . VMware version is VSPERE 5. There are 4-5 Vfilers running. When i force a failover from the one controler to other, the NFS clients have a delay about 30-40 seconds. Is this normal? Can i do something to investigate/tune the failover proccess? thnks in advance John
Hi John,
we have a FAS3240, too and the same behavior. I opened a Netapp Case (Case ID: 2002377504):
" During failover the IP address will be assigned to the surviving node, which does not reset. During giveback the filer reboots and this causes the network interface to reset. This is the timeout seen by the customer."
greets
Steffen
Von: toasters-bounces@teaparty.net [mailto:toasters-bounces@teaparty.net] Im Auftrag von John Sfetsoris Gesendet: Montag, 6. Februar 2012 11:21 An: toasters@teaparty.net Cc: jalterweb@yahoo.gr Betreff: Fas 3240 failover delay
Hello everybody
I have a Fas3204 HA in two seperate chassis , dataontap version 8.02P3 7mode. Many NFS servers are connected to it via iscsi connections. NFS servers are VMware mashines . VMware version is VSPERE 5. There are 4-5 Vfilers running. When i force a failover from the one controler to other, the NFS clients
have a delay about 30-40 seconds. Is this normal? Can i do something to investigate/tune the failover proccess?
thnks in advance
John
Hello John,
IIRC, the failover time can be up to 120 seconds which is why your connected systems need to have the correct disk and read timeout values set. For VMware ESX, you should install the vSphere Host utlities on the server which take care of the correct timeout values.
I'm not sure what you mean with "many nfs servers are connected via iscsi connections", since this sounds a bit odd.
Bye,
Alexander Griesser
System-Administrator
ANEXIA Internetdienstleistungs GmbH
Telefon: +43-463-208501-320
Telefax: +43-463-208501-500
E-Mail: ag@anexia.at mailto:ag@anexia.at
Web: http://www.anexia.at 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 John Sfetsoris Gesendet: Montag, 06. Februar 2012 11:21 An: toasters@teaparty.net Cc: jalterweb@yahoo.gr Betreff: Fas 3240 failover delay
Hello everybody
I have a Fas3204 HA in two seperate chassis , dataontap version 8.02P3 7mode. Many NFS servers are connected to it via iscsi connections. NFS servers are VMware mashines . VMware version is VSPERE 5. There are 4-5 Vfilers running. When i force a failover from the one controler to other, the NFS clients
have a delay about 30-40 seconds. Is this normal? Can i do something to investigate/tune the failover proccess?
thnks in advance
John
Check your network switches. You probably need to turn off SPANNING TREE. On a cisco, that could be: spanning-tree portfast enable (for a switch port) spanning-tree portfast trunk enable (for a trunk port) (check syntax for other switch vendors)
If you leave Spanning-tree on, there is about a 45 second delay before the port becomes ready.
--tmac Tim McCarthy Principal Consultant
RedHat Certified Engineer 804006984323821 (RHEL4) 805007643429572 (RHEL5)
On Mon, Feb 6, 2012 at 6:03 AM, Alexander Griesser ag@anexia.at wrote:
Hello John,
IIRC, the failover time can be up to 120 seconds which is why your connected systems need to have the correct disk and read timeout values set. For VMware ESX, you should install the vSphere Host utlities on the server which take care of the correct timeout values.
I’m not sure what you mean with „many nfs servers are connected via iscsi connections“, since this sounds a bit odd.
Bye,
Alexander Griesser
System-Administrator
ANEXIA Internetdienstleistungs GmbH
Telefon: +43-463-208501-320
Telefax: +43-463-208501-500
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 John Sfetsoris Gesendet: Montag, 06. Februar 2012 11:21
An: toasters@teaparty.net Cc: jalterweb@yahoo.gr Betreff: Fas 3240 failover delay
Hello everybody
I have a Fas3204 HA in two seperate chassis , dataontap version 8.02P3 7mode. Many NFS servers are connected to it via iscsi connections. NFS servers are VMware mashines . VMware version is VSPERE 5. There are 4-5 Vfilers running. When i force a failover from the one controler to other, the NFS clients
have a delay about 30-40 seconds. Is this normal? Can i do something to investigate/tune the failover proccess?
thnks in advance
John
Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
This will solve the problem you're having in about 95% of cases. Spanning-Tree can technically takes it's sweet time to allow the port to negotiate and online for up to 60 seconds if it chooses. If the behavior you're experiencing is that the takeover itself is fairly seamless (1-3 pings dropped), but the giveback takes a bit (more than 5 pings dropped) then spanning-tree is the culprit.
-----Original Message----- From: toasters-bounces@teaparty.net [mailto:toasters-bounces@teaparty.net] On Behalf Of tmac Sent: Monday, February 06, 2012 7:11 AM To: Alexander Griesser Cc: jalterweb@yahoo.gr; John Sfetsoris; toasters@teaparty.net Subject: Re: Fas 3240 failover delay
Check your network switches. You probably need to turn off SPANNING TREE. On a cisco, that could be: spanning-tree portfast enable (for a switch port) spanning-tree portfast trunk enable (for a trunk port) (check syntax for other switch vendors)
If you leave Spanning-tree on, there is about a 45 second delay before the port becomes ready.
--tmac Tim McCarthy Principal Consultant
RedHat Certified Engineer 804006984323821 (RHEL4) 805007643429572 (RHEL5)
On Mon, Feb 6, 2012 at 6:03 AM, Alexander Griesser ag@anexia.at wrote:
Hello John,
IIRC, the failover time can be up to 120 seconds which is why your connected systems need to have the correct disk and read timeout values set. For VMware ESX, you should install the vSphere Host utlities on the server which take care of the correct timeout values.
I'm not sure what you mean with "many nfs servers are connected via iscsi connections", since this sounds a bit odd.
Bye,
Alexander Griesser
System-Administrator
ANEXIA Internetdienstleistungs GmbH
Telefon: +43-463-208501-320
Telefax: +43-463-208501-500
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 John Sfetsoris Gesendet: Montag, 06. Februar 2012 11:21
An: toasters@teaparty.net Cc: jalterweb@yahoo.gr Betreff: Fas 3240 failover delay
Hello everybody
I have a Fas3204 HA in two seperate chassis , dataontap version 8.02P3 7mode. Many NFS servers are connected to it via iscsi connections. NFS servers are VMware mashines . VMware version is VSPERE 5. There are 4-5 Vfilers running. When i force a failover from the one controler to other, the NFS clients
have a delay about 30-40 seconds. Is this normal? Can i do something to investigate/tune the failover proccess?
thnks in advance
John
Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
_______________________________________________ Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
Hi Tim,
you're the man of the week ;) It's a shame, but i forgot to check this. Thanks for your help!
greets
Steffen
-----Ursprüngliche Nachricht----- Von: toasters-bounces@teaparty.net [mailto:toasters-bounces@teaparty.net] Im Auftrag von tmac Gesendet: Montag, 6. Februar 2012 14:11 An: Alexander Griesser Cc: jalterweb@yahoo.gr; John Sfetsoris; toasters@teaparty.net Betreff: Re: Fas 3240 failover delay
Check your network switches. You probably need to turn off SPANNING TREE. On a cisco, that could be: spanning-tree portfast enable (for a switch port) spanning-tree portfast trunk enable (for a trunk port) (check syntax for other switch vendors)
If you leave Spanning-tree on, there is about a 45 second delay before the port becomes ready.
--tmac Tim McCarthy Principal Consultant
RedHat Certified Engineer 804006984323821 (RHEL4) 805007643429572 (RHEL5)
Thnk you all for your answers. They are all very helpfull The problem i am facing occurs from the time the one node falls and the surving node starts the take over procces until the time the take over proccess completes and the node is ready to serve data. This time is about 30-40 seconds
________________________________________ From: toasters-bounces@teaparty.net [toasters-bounces@teaparty.net] on behalf of Steffen Knauf [sknauf@chipxonio.de] Sent: Thursday, February 09, 2012 11:36 AM To: 'tmac' Cc: toasters@teaparty.net Subject: AW: Fas 3240 failover delay
Hi Tim,
you're the man of the week ;) It's a shame, but i forgot to check this. Thanks for your help!
greets
Steffen
-----Ursprüngliche Nachricht----- Von: toasters-bounces@teaparty.net [mailto:toasters-bounces@teaparty.net] Im Auftrag von tmac Gesendet: Montag, 6. Februar 2012 14:11 An: Alexander Griesser Cc: jalterweb@yahoo.gr; John Sfetsoris; toasters@teaparty.net Betreff: Re: Fas 3240 failover delay
Check your network switches. You probably need to turn off SPANNING TREE. On a cisco, that could be: spanning-tree portfast enable (for a switch port) spanning-tree portfast trunk enable (for a trunk port) (check syntax for other switch vendors)
If you leave Spanning-tree on, there is about a 45 second delay before the port becomes ready.
--tmac Tim McCarthy Principal Consultant
RedHat Certified Engineer 804006984323821 (RHEL4) 805007643429572 (RHEL5)
_______________________________________________ Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters