using fcalias is not the best practices, device aliases is. Also please consider talking about SmartZoning as that negates some of the best practices your provided above.
Thanks for reading the blog carefully and raising your opinion.
We are aware that device alias is a fabric wide feature and more convenient than fcalias. However, considering multiple bugs for device alias found in NX-OS v5.x working with DCNM, we still recommend our customer to use fcalias.
Smart zoning is a new feature introduced in 5.2(6) which saves a lot of time for SAN administrators to configure zoning.
We encourage customer to use this feature and have added some recommend into this article.
Thanks again for providing your valuable suggestion.
can you please point me to Cisco bug reports about device-aliases. We have been using devices aliases for at least 3 years and i have yet to find any issues.
majority of those are bugs in really old 5.2.x branch of NX-OS. Folks should be on the latest 5.2.8 if only for the reason to address ShellShock vulnerabilities on NX-OS. These should not be a deterrent from using device aliases. Being able to see server name when you create IVR zones, when you run sh flogi database and others is just too great of an feature to be not considered.
Below zone configuration is correct way configuration? I have noticed that one fabric it was configured with WWPN and other WWNN. Please help me on this. how to correct this? HBA1 Port WWN:10:00:00:00:c9:e1:87:1d Node WWN:20:00:00:00:c9:e1:87:1d HBA2 Port WWN: 10:00:00:00:c9:b9:52:db Node WWN: 20:00:00:00:c9:b9:52:db Switch A> zone: z_WinHostSQL01_HBA1_SANCONB_1B 50:0a:09:82:9d:c9:f1:5c 10:00:00:00:c9:e1:87:1d zone: z_WinHostSQL01_HBA1_SANCONA_0C 50:0a:09:83:8d:c9:f1:5c 10:00:00:00:c9:e1:87:1d zone: z_WinHostSQL01_HBA1_SANCONA_1A 50:0a:09:81:8d:c9:f1:5c 10:00:00:00:c9:e1:87:1d zone: z_WinHostSQL01_HBA1_SANCONB_0D 50:0a:09:84:9d:c9:f1:5c Switch A> Switch B> zone: z_WinHostSQL01_HBA2_SANCONA_0D 50:0a:09:84:8d:c9:f1:5c 20:00:00:00:c9:b9:52:db zone: z_WinHostSQL01_HBA2_SANCONA_1B 50:0a:09:82:8d:c9:f1:5c 20:00:00:00:c9:b9:52:db zone: z_WinHostSQL01_HBA2_SANCONB_0C 50:0a:09:83:9d:c9:f1:5c 20:00:00:00:c9:b9:52:db zone: z_WinHostSQL01_HBA2_SANCONB_1A 50:0a:09:83:9d:c9:f1:5c 20:00:00:00:c9:b9:52:db Switch B>
If your HBAs are all single port ones, your current configuraiton should work. But once you add a dual port HBA to your system, you must upgdate the config.
Sorry if i am asking dumb question.i am new to this zoning concept.
Below is the my previous setup in both fabrics. HBA1 connected to one fabric with (PWWN). second HBA2 connected to other fabric(NWWN) as per my organization standards.
So if we keep the same configuration when ever switch takeover happened fro SwitchA to SwitchB all my mapped luns will accessible right from SwitchB?
if Yes.i simply worrying with existing concept. so i will revert back my last thread.
FYI Note: Switch uptime is 435 days. this zoning was done by year ago.we got this project recently. out of 7 windows nodes 4 was zoning configured by PWWN in both fabrics and remaining 3 are configured as below one fabric pwwn and other nwwn.
this is not an issue ?
HBA1
Port WWN:10:00:00:00:c9:e1:87:1d
Node WWN:20:00:00:00:c9:e1:87:1d
HBA2
Port WWN: 10:00:00:00:c9:b9:52:db
Node WWN: 20:00:00:00:c9:b9:52:db
Switch A>
Effective configuration: CFS_SWITCHA
zone: z_WinHostSQL01_HBA1_SANCONB_1B
50:0a:09:82:9d:c9:f1:5c
10:00:00:00:c9:e1:87:1d
zone: z_WinHostSQL01_HBA1_SANCONA_0C
50:0a:09:83:8d:c9:f1:5c
10:00:00:00:c9:e1:87:1d
zone: z_WinHostSQL01_HBA1_SANCONA_1A
50:0a:09:81:8d:c9:f1:5c
10:00:00:00:c9:e1:87:1d
zone: z_WinHostSQL01_HBA1_SANCONB_0D
50:0a:09:84:9d:c9:f1:5c
10:00:00:00:c9:e1:87:1d
Switch A>
Switch B>
Effective configuration: CFS_SWITCHB
zone: z_WinHostSQL01_HBA2_SANCONA_0D
50:0a:09:84:8d:c9:f1:5c
20:00:00:00:c9:b9:52:db
zone: z_WinHostSQL01_HBA2_SANCONA_1B
50:0a:09:82:8d:c9:f1:5c
20:00:00:00:c9:b9:52:db
zone: z_WinHostSQL01_HBA2_SANCONB_0C
50:0a:09:83:9d:c9:f1:5c
20:00:00:00:c9:b9:52:db
zone: z_WinHostSQL01_HBA2_SANCONB_1A
50:0a:09:83:9d:c9:f1:5c
20:00:00:00:c9:b9:52:db
Switch B>
Please bare me and provide the simply reply. thanks. God Bless you..!
there is no "takeover". Your host should be zoned to the array using Switch A and Switch B. Both switches are active at the same time, both HBAs on the host are active at the same time.
Your hosts gets zoned to multiple ports on the storage array so you need to ensure that host is running some sort of multipathing software to ensure in case of switch/hba/cable failure, your host will continue to access the storage through available paths.
dynamox
9 Legend
•
20.4K Posts
0
January 20th, 2015 10:00
using fcalias is not the best practices, device aliases is. Also please consider talking about SmartZoning as that negates some of the best practices your provided above.
ECN-APJ
2 Intern
•
308 Posts
0
January 30th, 2015 01:00
Hi dynamox,
Thanks for reading the blog carefully and raising your opinion.
We are aware that device alias is a fabric wide feature and more convenient than fcalias. However, considering multiple bugs for device alias found in NX-OS v5.x working with DCNM, we still recommend our customer to use fcalias.
Smart zoning is a new feature introduced in 5.2(6) which saves a lot of time for SAN administrators to configure zoning.
We encourage customer to use this feature and have added some recommend into this article.
Thanks again for providing your valuable suggestion.
dynamox
9 Legend
•
20.4K Posts
0
January 30th, 2015 02:00
can you please point me to Cisco bug reports about device-aliases. We have been using devices aliases for at least 3 years and i have yet to find any issues.
Thanks
ECN-APJ
2 Intern
•
308 Posts
0
January 31st, 2015 02:00
Hi Dynamox,
Here is a brief list of the related defects.
CSCtu03947 - Device-alias add/delete doesn't update client cache right away.
Affecting: 5.2.1, 5.2.2
Fixed in: 5.2(2)S71, 5.2(2.72)S0, 5.2(2a)S3, 7.0(1)ZD(0.3)
CSCtt20652 - DCNM SAN: Device Alias CFS Regions column does not sort
Affecting: 5.2.1
Fixed in: 6.2(0.18)S0
CSCuc04839 - Device-alias instead of fcalias in the document under DCNM control panel
Affecting: 5.2(7.9)
Fixed in: Release Pending
CSCtn77734 - FM Cannot purge deleted device-aliases they show as decommissioned
Affecting: 5.0(4a)
Fixed in: 5.0(4c)S2, 5.0(6.54)S
dynamox
9 Legend
•
20.4K Posts
0
January 31st, 2015 06:00
majority of those are bugs in really old 5.2.x branch of NX-OS. Folks should be on the latest 5.2.8 if only for the reason to address ShellShock vulnerabilities on NX-OS. These should not be a deterrent from using device aliases. Being able to see server name when you create IVR zones, when you run sh flogi database and others is just too great of an feature to be not considered.
Teja1988
5 Posts
0
November 13th, 2015 20:00
Below zone configuration is correct way configuration? I have noticed that one fabric it was configured with WWPN and other WWNN. Please help me on this. how to correct this? HBA1 Port WWN:10:00:00:00:c9:e1:87:1d Node WWN:20:00:00:00:c9:e1:87:1d HBA2 Port WWN: 10:00:00:00:c9:b9:52:db Node WWN: 20:00:00:00:c9:b9:52:db Switch A> zone: z_WinHostSQL01_HBA1_SANCONB_1B 50:0a:09:82:9d:c9:f1:5c 10:00:00:00:c9:e1:87:1d zone: z_WinHostSQL01_HBA1_SANCONA_0C 50:0a:09:83:8d:c9:f1:5c 10:00:00:00:c9:e1:87:1d zone: z_WinHostSQL01_HBA1_SANCONA_1A 50:0a:09:81:8d:c9:f1:5c 10:00:00:00:c9:e1:87:1d zone: z_WinHostSQL01_HBA1_SANCONB_0D 50:0a:09:84:9d:c9:f1:5c Switch A> Switch B> zone: z_WinHostSQL01_HBA2_SANCONA_0D 50:0a:09:84:8d:c9:f1:5c 20:00:00:00:c9:b9:52:db zone: z_WinHostSQL01_HBA2_SANCONA_1B 50:0a:09:82:8d:c9:f1:5c 20:00:00:00:c9:b9:52:db zone: z_WinHostSQL01_HBA2_SANCONB_0C 50:0a:09:83:9d:c9:f1:5c 20:00:00:00:c9:b9:52:db zone: z_WinHostSQL01_HBA2_SANCONB_1A 50:0a:09:83:9d:c9:f1:5c 20:00:00:00:c9:b9:52:db Switch B>
dynamox
9 Legend
•
20.4K Posts
0
November 15th, 2015 12:00
zoning needs to be done using PWWN
ECN-APJ
2 Intern
•
308 Posts
0
November 17th, 2015 19:00
Hi Teja1988,
Agree with dynamox, EMC recommends Soft/WWPN/pWWN zoning (over Hard/Port zoning).
Teja1988
5 Posts
0
November 18th, 2015 07:00
Hi I have performed the step 2 on second fiber. Please check and update me the same.
stpe 1)
Switch B>
aliadd "WinHostSQL01_HBA2", "10:00:00:00:c9:b9:52:db"
aliremove "WinHostSQL01_HBA2", "20:00:00:00:c9:b9:52:db"
cfgsave
cfgenable "cfg_file"
Switch B>
Step 2)
Switch B>
aliadd "WinHostSQL01_HBA2", "10:00:00:00:c9:b9:52:db"
cfgsave
cfgenable "cfg_file"
Switch B>
I have just added new ali "WWPN" to the existing zone. It is okay to be put like this? any performance issue occur keeping WWPN and WWNN in same zone.
SwitchB>
zone: WinHostSQL01_HBA2_SANCONA_0D
50:0a:09:84:8d:c9:f1:5c
20:00:00:00:c9:b9:52:db
10:00:00:00:c9:b9:52:db
zone: WinHostSQL01_HBA2_SANCONA_1B
50:0a:09:82:8d:c9:f1:5c
20:00:00:00:c9:b9:52:db
10:00:00:00:c9:b9:52:db
zone: WinHostSQL01_HBA2_SANCONB_0C
50:0a:09:83:9d:c9:f1:5c
20:00:00:00:c9:b9:52:db
10:00:00:00:c9:b9:52:db
zone: WinHostSQL01_HBA2_SANCONB_1A
50:0a:09:83:9d:c9:f1:5c
20:00:00:00:c9:b9:52:db
10:00:00:00:c9:b9:52:db
SwithB>
ECN-APJ
2 Intern
•
308 Posts
0
November 18th, 2015 21:00
A single port HBA will have WWNN & WWPN as same.
A dual port HBA will have 1 WWNN & 2 WWPNs.
If your HBAs are all single port ones, your current configuraiton should work. But once you add a dual port HBA to your system, you must upgdate the config.
Teja1988
5 Posts
0
November 19th, 2015 08:00
Sorry if i am asking dumb question.i am new to this zoning concept.
Below is the my previous setup in both fabrics. HBA1 connected to one fabric with (PWWN). second HBA2 connected to other fabric(NWWN) as per my organization standards.
So if we keep the same configuration when ever switch takeover happened fro SwitchA to SwitchB all my mapped luns will accessible right from SwitchB?
if Yes.i simply worrying with existing concept. so i will revert back my last thread.
FYI Note: Switch uptime is 435 days. this zoning was done by year ago.we got this project recently. out of 7 windows nodes 4 was zoning configured by PWWN in both fabrics and remaining 3 are configured as below one fabric pwwn and other nwwn.
this is not an issue ?
HBA1
Port WWN:10:00:00:00:c9:e1:87:1d
Node WWN:20:00:00:00:c9:e1:87:1d
HBA2
Port WWN: 10:00:00:00:c9:b9:52:db
Node WWN: 20:00:00:00:c9:b9:52:db
Switch A>
Effective configuration: CFS_SWITCHA
zone: z_WinHostSQL01_HBA1_SANCONB_1B
50:0a:09:82:9d:c9:f1:5c
10:00:00:00:c9:e1:87:1d
zone: z_WinHostSQL01_HBA1_SANCONA_0C
50:0a:09:83:8d:c9:f1:5c
10:00:00:00:c9:e1:87:1d
zone: z_WinHostSQL01_HBA1_SANCONA_1A
50:0a:09:81:8d:c9:f1:5c
10:00:00:00:c9:e1:87:1d
zone: z_WinHostSQL01_HBA1_SANCONB_0D
50:0a:09:84:9d:c9:f1:5c
10:00:00:00:c9:e1:87:1d
Switch A>
Switch B>
Effective configuration: CFS_SWITCHB
zone: z_WinHostSQL01_HBA2_SANCONA_0D
50:0a:09:84:8d:c9:f1:5c
20:00:00:00:c9:b9:52:db
zone: z_WinHostSQL01_HBA2_SANCONA_1B
50:0a:09:82:8d:c9:f1:5c
20:00:00:00:c9:b9:52:db
zone: z_WinHostSQL01_HBA2_SANCONB_0C
50:0a:09:83:9d:c9:f1:5c
20:00:00:00:c9:b9:52:db
zone: z_WinHostSQL01_HBA2_SANCONB_1A
50:0a:09:83:9d:c9:f1:5c
20:00:00:00:c9:b9:52:db
Switch B>
Please bare me and provide the simply reply. thanks. God Bless you..!
Thanks,
Teja
dynamox
9 Legend
•
20.4K Posts
0
November 22nd, 2015 06:00
there is no "takeover". Your host should be zoned to the array using Switch A and Switch B. Both switches are active at the same time, both HBAs on the host are active at the same time.
Your hosts gets zoned to multiple ports on the storage array so you need to ensure that host is running some sort of multipathing software to ensure in case of switch/hba/cable failure, your host will continue to access the storage through available paths.