search

Seagate / propeller

Customized lvmlockd
GNU General Public License v2.0
4 stars 7 forks source link

Different drives share the same firmware? #58

Closed Leo-Yan closed 4 years ago

Leo-Yan commented 4 years ago

Hi @TProhofsky ,

When tested the code and found below behavior:

Vendor   Handle       Model Number            Serial Number          FwRev
QEMU     /dev/sg0     QEMU DVD-ROM                QEMU    QEMU DVD   1.5. 
IET      /dev/sg1     Controller                                     0001 
SEAGATE  /dev/sg2     XS3840SE70014           HLL03109               1022 
SEAGATE  /dev/sg3     XS3840SE70014           HLL030ZP               1022 
SEAGATE  /dev/sg4     XS3840SE70014           HLL03102               1022 
SEAGATE  /dev/sg5     XS3840SE70014           HLL0310Z               1022 
SEAGATE  /dev/sg6     XS3840SE70014           HLL030ZP               1022 
SEAGATE  /dev/sg7     XS3840SE70014           HLL03102               1022 
SEAGATE  /dev/sg8     XS3840SE70014           HLL0310Z               1022 
SEAGATE  /dev/sg9     XS3840SE70014           HLL03109               1022

In the system, the device /dev/sg2 and /dev/sg9 have the same model number and serial number; and if send IDM SCSI command to sg2 to acquire mutex, and I can confirm that sg9 contains the exactly same mutex content.

So can I get conclusion that sg2 and sg9 share the same one firmware copy? If so, should we send command to only one of (sg2 and sg9) so can avoid duplicate acquiring mutex?

TProhofsky commented 4 years ago

Leo,

Those duplicate SCSI handles are the same drive with both ports connected. All actions will behave the same using either port to access the drive. Have Glen disconnect one of the cables to start with a single path configuration.

-=Tom

From: Leo Yan notifications@github.com Sent: Monday, June 8, 2020 1:12 AM To: Seagate/propeller propeller@noreply.github.com Cc: Tom Prohofsky tom.r.prohofsky@seagate.com; Mention mention@noreply.github.com Subject: [Seagate/propeller] Different drives share the same firmware? (#58)

Hi @TProhofskyhttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_TProhofsky&d=DwMCaQ&c=IGDlg0lD0b-nebmJJ0Kp8A&r=hkffANeUJqcNQ6_JRn1xH-8Iaq-39QwpGZFZg6BME-U&m=OQSfehfwoBZlgAvMN7r4LD17tFQfpQMmsx1yA3NMaMc&s=VJvvSOLYZS4dzCJUYF-5axFtOLJimnzgjAvx_Az9KUQ&e= ,

When tested the code and found below behavior:

Vendor Handle Model Number Serial Number FwRev QEMU /dev/sg0 QEMU DVD-ROM QEMU QEMU DVD 1.5. IET /dev/sg1 Controller 0001 SEAGATE /dev/sg2 XS3840SE70014 HLL03109 1022 SEAGATE /dev/sg3 XS3840SE70014 HLL030ZP 1022 SEAGATE /dev/sg4 XS3840SE70014 HLL03102 1022 SEAGATE /dev/sg5 XS3840SE70014 HLL0310Z 1022 SEAGATE /dev/sg6 XS3840SE70014 HLL030ZP 1022 SEAGATE /dev/sg7 XS3840SE70014 HLL03102 1022 SEAGATE /dev/sg8 XS3840SE70014 HLL0310Z 1022 SEAGATE /dev/sg9 XS3840SE70014 HLL03109 1022

In the system, the device /dev/sg2 and /dev/sg9 have the same model number and serial number; and if send IDM SCSI command to sg2 to acquire mutex, and I can confirm that sg9 contains the exactly same mutex content.

So can I get conclusion that sg2 and sg9 share the same one firmware copy? If so, should we send command to only one of (sg2 and sg9) so can avoid duplicate acquiring mutex?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_Seagate_propeller_issues_58&d=DwMCaQ&c=IGDlg0lD0b-nebmJJ0Kp8A&r=hkffANeUJqcNQ6_JRn1xH-8Iaq-39QwpGZFZg6BME-U&m=OQSfehfwoBZlgAvMN7r4LD17tFQfpQMmsx1yA3NMaMc&s=VHMHSoCdVSbkBtqppvhOT8dQF1K0YlHtqrFMx37l9KU&e=, or unsubscribehttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_AJT56UAIOXPSSSLZBRZ3HHLRVR6MFANCNFSM4NYA7ESQ&d=DwMCaQ&c=IGDlg0lD0b-nebmJJ0Kp8A&r=hkffANeUJqcNQ6_JRn1xH-8Iaq-39QwpGZFZg6BME-U&m=OQSfehfwoBZlgAvMN7r4LD17tFQfpQMmsx1yA3NMaMc&s=ZCBHH0ULosgT0JeMAPdCGEjxku59-_PRCZPEFNpyjUo&e=.

Leo-Yan commented 4 years ago

Thanks for confirmation, Tom.

As the first trying, I don't need to ask Glen to disconnect one of cable, alternatively, I can use ISCSI target to only export half of SCSI devices to be used. This is more flexible and I will add code to try to handle for this duplicate case.

TProhofsky commented 4 years ago

Please use serial number for the index key in your design where possible. 1 or more SCSI devices may over multiple paths to the drive. I'll open a github issue to track the design approach for multipath.

-=Tom

From: Leo Yan notifications@github.com Sent: Monday, June 8, 2020 9:34 AM To: Seagate/propeller propeller@noreply.github.com Cc: Tom Prohofsky tom.r.prohofsky@seagate.com; Mention mention@noreply.github.com Subject: Re: [Seagate/propeller] Different drives share the same firmware? (#58)

Thanks for confirmation, Tom.

As the first trying, I don't need to ask Glen to disconnect one of cable, alternatively, I can use ISCSI target to only export half of SCSI devices to be used. This is more flexible and I will add code to try to handle for this duplicate case.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_Seagate_propeller_issues_58-23issuecomment-2D640667080&d=DwMCaQ&c=IGDlg0lD0b-nebmJJ0Kp8A&r=hkffANeUJqcNQ6_JRn1xH-8Iaq-39QwpGZFZg6BME-U&m=MzjUJ8-e0TbWWgxDVN01phwA9GoAyojD0hMhiyQX8m8&s=ycyT0sYczAYnSNebrd8P6wYiUKorYvEFJBtk7zUD_Rc&e=, or unsubscribehttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_AJT56UEULXLO3ECEDA2WEH3RVTZJHANCNFSM4NYA7ESQ&d=DwMCaQ&c=IGDlg0lD0b-nebmJJ0Kp8A&r=hkffANeUJqcNQ6_JRn1xH-8Iaq-39QwpGZFZg6BME-U&m=MzjUJ8-e0TbWWgxDVN01phwA9GoAyojD0hMhiyQX8m8&s=xQ18ugpwZR30yQR3JsRzPJbmG-U_5MSw0TqaoqqUhMA&e=.

Leo-Yan commented 4 years ago

Let's use ticket https://github.com/Seagate/propeller/issues/59 to track multi path related discussion. For parsing the duplicate paths in device list, which has been fixed in the latest IDM lock manager.