Hi all,
We are getting errors while trying to replace with a new HP LTO3 drive, the actual LTO1 drive connected to the scsi adapter X2027A of a FAS920 NetApp Release 7.0.4 (see errors and actual configuration below).
NetApp Log # 2475693
I would really appreciate any hints..
Actual Configuration
Model Name: FAS920 NetApp Release 7.0.4
LTO1 drive connected to scsi Adapter
slot 10: SCSI Host Adapter 10a (Adaptec AIC7902, <ONLINE>)
HW Version 0.0
Ultra320 SCSI, Low Voltage Differential
I/O base 0xdd00, size 0x100
memory mapped I/O base 0xa04c0000, size 0x2000
I/O base 0xdf00, size 0x100
slot 10: SCSI Host Adapter 10b (Adaptec AIC7902, <ONLINE>)
HW Version 0.0
Ultra320 SCSI, Low Voltage Differential
4: Tape: Seagate ULTRIUM06242 1613
I/O base 0xde00, size 0x100
memory mapped I/O base 0xa04c2000, size 0x2000
I/O base 0xdc00, size 0x100
SCSI ERRORS
Sat Sep 8 11:04:38 VET [adpch_timeoutd:warning]: adpch_timeoutd: 10a.3(0x0568ede0, 0x12): command timeout, quiescing bus to allow outstanding I/O to complete(ticks = 8). Sat Sep 8 11:04:46 VET [adpch_asyncd:warning]: adpch asyncd: Resetting SCSI 10a target 3 to clear outstanding I/O. Sat Sep 8 11:04:46 VET [adpch_intrd:warning]: adpch_timeoutd: 10a: unquiescing bus Sat Sep 8 11:04:57 VET [adpch_timeoutd:warning]: adpch_timeoutd: 10a.3(0x0568ede0, 0x12): command timeout, quiescing bus to allow outstanding I/O to complete(ticks = 8). Sat Sep 8 11:05:05 VET [adpch_asyncd:warning]: adpch asyncd: Resetting SCSI 10a target 3 to clear outstanding I/O. Sat Sep 8 11:05:05 VET [adpch_intrd:warning]: adpch_timeoutd: 10a: unquiescing bus Sat Sep 8 11:06:10 VET [adpch_intrd:warning]: Unexpected page code: Ultrium 3 serial number not available. Sat Sep 8 11:34:49 VET [rc:info]: relog syslog Sat Sep 8 11:33:24 VET [adpch_asyncd:warning]: adpch asyncd: Resetting SCSI bus 10a to clear outstanding I/O. Sat Sep 8 11:34:56 VET [adpch_asyncd:warning]: adpch asyncd: Resetting SCSI bus 10a to clear outstanding I/O. Sat Sep 8 11:36:28 VET [adpch_asyncd:warning]: adpch asyncd: Resetting SCSI bus 10a to clear outstanding I/O. Sat Sep 8 11:41:17 VET [adpch_asyncd:warning]: adpch asyncd: Resetting SCSI bus 10b to clear outstanding I/O.
Check the SCSI connections to: 1. make sure ther are no little bent pins 2. screwed down TIGHT
If problems persist, try a different cable. Problems still, RMA the SCSI card. Lastly, RMA the Tape drive itself.
--tmac
On 9/19/07, Ernesto Silvera esilvera@csrven.com wrote:
Hi all,
We are getting errors while trying to replace with a new HP LTO3 drive, the actual LTO1 drive connected to the scsi adapter X2027A of a FAS920 NetApp Release 7.0.4 (see errors and actual configuration below).
NetApp Log # 2475693
I would really appreciate any hints….
Actual Configuration
Model Name: FAS920 NetApp Release 7.0.4
LTO1 drive connected to scsi Adapter
slot 10: SCSI Host Adapter 10a (Adaptec AIC7902, <ONLINE>)
HW Version 0.0 Ultra320 SCSI, Low Voltage Differential I/O base 0xdd00, size 0x100 memory mapped I/O base 0xa04c0000, size 0x2000 I/O base 0xdf00, size 0x100 slot 10: SCSI Host Adapter 10b (Adaptec AIC7902, <ONLINE>) HW Version 0.0 Ultra320 SCSI, Low Voltage Differential 4: Tape: Seagate ULTRIUM06242 1613 I/O base 0xde00, size 0x100 memory mapped I/O base 0xa04c2000, size 0x2000 I/O base 0xdc00, size 0x100SCSI ERRORS
Sat Sep 8 11:04:38 VET [adpch_timeoutd:warning]: adpch_timeoutd: 10a.3(0x0568ede0, 0x12): command timeout, quiescing bus to allow outstanding I/O to complete(ticks = 8). Sat Sep 8 11:04:46 VET [adpch_asyncd:warning]: adpch asyncd: Resetting SCSI 10a target 3 to clear outstanding I/O. Sat Sep 8 11:04:46 VET [adpch_intrd:warning]: adpch_timeoutd: 10a: unquiescing bus Sat Sep 8 11:04:57 VET [adpch_timeoutd:warning]: adpch_timeoutd: 10a.3(0x0568ede0, 0x12): command timeout, quiescing bus to allow outstanding I/O to complete(ticks = 8). Sat Sep 8 11:05:05 VET [adpch_asyncd:warning]: adpch asyncd: Resetting SCSI 10a target 3 to clear outstanding I/O. Sat Sep 8 11:05:05 VET [adpch_intrd:warning]: adpch_timeoutd: 10a: unquiescing bus Sat Sep 8 11:06:10 VET [adpch_intrd:warning]: Unexpected page code: Ultrium 3 serial number not available. Sat Sep 8 11:34:49 VET [rc:info]: relog syslog Sat Sep 8 11:33:24 VET [adpch_asyncd:warning]: adpch asyncd: Resetting SCSI bus 10a to clear outstanding I/O. Sat Sep 8 11:34:56 VET [adpch_asyncd:warning]: adpch asyncd: Resetting SCSI bus 10a to clear outstanding I/O. Sat Sep 8 11:36:28 VET [adpch_asyncd:warning]: adpch asyncd: Resetting SCSI bus 10a to clear outstanding I/O. Sat Sep 8 11:41:17 VET [adpch_asyncd:warning]: adpch asyncd: Resetting SCSI bus 10b to clear outstanding I/O.
We already try another cable, drive and terminator...
Netapp replies that we are hitting the Bug 153964, "SCSI transfers greater than 64KB hang on X2027A in Ultra320 mode". So the question is what is the blocksize Netapp uses for the dump?. It is 64KB?
Regards!
-----Original Message----- From: tmac [mailto:tmacmd@gmail.com] Sent: Wednesday, September 19, 2007 8:17 AM To: Ernesto Silvera Cc: toasters@mathworks.com Subject: Re: SCSI LTO3 Drive for NDMP Backup problems
Check the SCSI connections to: 1. make sure ther are no little bent pins 2. screwed down TIGHT
If problems persist, try a different cable. Problems still, RMA the SCSI card. Lastly, RMA the Tape drive itself.
--tmac
On 9/19/07, Ernesto Silvera esilvera@csrven.com wrote:
Hi all,
We are getting errors while trying to replace with a new HP LTO3 drive,
the
actual LTO1 drive connected to the scsi adapter X2027A of a FAS920 NetApp Release 7.0.4 (see errors and actual configuration below).
NetApp Log # 2475693
I would really appreciate any hints..
Actual Configuration
Model Name: FAS920 NetApp Release 7.0.4
LTO1 drive connected to scsi Adapter
slot 10: SCSI Host Adapter 10a (Adaptec AIC7902, <ONLINE>)
HW Version 0.0 Ultra320 SCSI, Low Voltage Differential I/O base 0xdd00, size 0x100 memory mapped I/O base 0xa04c0000, size 0x2000 I/O base 0xdf00, size 0x100 slot 10: SCSI Host Adapter 10b (Adaptec AIC7902, <ONLINE>) HW Version 0.0 Ultra320 SCSI, Low Voltage Differential 4: Tape: Seagate ULTRIUM06242 1613 I/O base 0xde00, size 0x100 memory mapped I/O base 0xa04c2000, size 0x2000 I/O base 0xdc00, size 0x100SCSI ERRORS
Sat Sep 8 11:04:38 VET [adpch_timeoutd:warning]: adpch_timeoutd: 10a.3(0x0568ede0, 0x12): command timeout, quiescing bus to allow
outstanding
I/O to complete(ticks = 8). Sat Sep 8 11:04:46 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
10a target 3 to clear outstanding I/O. Sat Sep 8 11:04:46 VET [adpch_intrd:warning]: adpch_timeoutd: 10a: unquiescing bus Sat Sep 8 11:04:57 VET [adpch_timeoutd:warning]: adpch_timeoutd: 10a.3(0x0568ede0, 0x12): command timeout, quiescing bus to allow
outstanding
I/O to complete(ticks = 8). Sat Sep 8 11:05:05 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
10a target 3 to clear outstanding I/O. Sat Sep 8 11:05:05 VET [adpch_intrd:warning]: adpch_timeoutd: 10a: unquiescing bus Sat Sep 8 11:06:10 VET [adpch_intrd:warning]: Unexpected page code:
Ultrium
3 serial number not available. Sat Sep 8 11:34:49 VET [rc:info]: relog syslog Sat Sep 8 11:33:24 VET [adpch_asyncd:warning]: adpch asyncd: Resetting SCSI bus 10a to clear outstanding I/O. Sat Sep 8 11:34:56 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
bus 10a to clear outstanding I/O. Sat Sep 8 11:36:28 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
bus 10a to clear outstanding I/O. Sat Sep 8 11:41:17 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
bus 10b to clear outstanding I/O.
From the bug ID you supplied:
Check vendor documentation for SCSI speed capabilities. If the device is capable of Ultra-320 speeds, restrict NDMP and dump/restore record sizes to 64KB or smaller when the drives are attached via the X2027A SCSI card. There is no workaround for snapmirror.
How are you doing dumps? If via NDMP you need to talk with your software vendor to figure how to change the buffer size down a little (like to 32K or something).
--tmac
On 9/19/07, Ernesto Silvera esilvera@csrven.com wrote:
We already try another cable, drive and terminator...
Netapp replies that we are hitting the Bug 153964, "SCSI transfers greater than 64KB hang on X2027A in Ultra320 mode". So the question is what is the blocksize Netapp uses for the dump?. It is 64KB?
Regards!
-----Original Message----- From: tmac [mailto:tmacmd@gmail.com] Sent: Wednesday, September 19, 2007 8:17 AM To: Ernesto Silvera Cc: toasters@mathworks.com Subject: Re: SCSI LTO3 Drive for NDMP Backup problems
Check the SCSI connections to:
- make sure ther are no little bent pins
- screwed down TIGHT
If problems persist, try a different cable. Problems still, RMA the SCSI card. Lastly, RMA the Tape drive itself.
--tmac
On 9/19/07, Ernesto Silvera esilvera@csrven.com wrote:
Hi all,
We are getting errors while trying to replace with a new HP LTO3 drive,
the
actual LTO1 drive connected to the scsi adapter X2027A of a FAS920 NetApp Release 7.0.4 (see errors and actual configuration below).
NetApp Log # 2475693
I would really appreciate any hints..
Actual Configuration
Model Name: FAS920 NetApp Release 7.0.4
LTO1 drive connected to scsi Adapter
slot 10: SCSI Host Adapter 10a (Adaptec AIC7902, <ONLINE>)
HW Version 0.0 Ultra320 SCSI, Low Voltage Differential I/O base 0xdd00, size 0x100 memory mapped I/O base 0xa04c0000, size 0x2000 I/O base 0xdf00, size 0x100 slot 10: SCSI Host Adapter 10b (Adaptec AIC7902, <ONLINE>) HW Version 0.0 Ultra320 SCSI, Low Voltage Differential 4: Tape: Seagate ULTRIUM06242 1613 I/O base 0xde00, size 0x100 memory mapped I/O base 0xa04c2000, size 0x2000 I/O base 0xdc00, size 0x100SCSI ERRORS
Sat Sep 8 11:04:38 VET [adpch_timeoutd:warning]: adpch_timeoutd: 10a.3(0x0568ede0, 0x12): command timeout, quiescing bus to allow
outstanding
I/O to complete(ticks = 8). Sat Sep 8 11:04:46 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
10a target 3 to clear outstanding I/O. Sat Sep 8 11:04:46 VET [adpch_intrd:warning]: adpch_timeoutd: 10a: unquiescing bus Sat Sep 8 11:04:57 VET [adpch_timeoutd:warning]: adpch_timeoutd: 10a.3(0x0568ede0, 0x12): command timeout, quiescing bus to allow
outstanding
I/O to complete(ticks = 8). Sat Sep 8 11:05:05 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
10a target 3 to clear outstanding I/O. Sat Sep 8 11:05:05 VET [adpch_intrd:warning]: adpch_timeoutd: 10a: unquiescing bus Sat Sep 8 11:06:10 VET [adpch_intrd:warning]: Unexpected page code:
Ultrium
3 serial number not available. Sat Sep 8 11:34:49 VET [rc:info]: relog syslog Sat Sep 8 11:33:24 VET [adpch_asyncd:warning]: adpch asyncd: Resetting SCSI bus 10a to clear outstanding I/O. Sat Sep 8 11:34:56 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
bus 10a to clear outstanding I/O. Sat Sep 8 11:36:28 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
bus 10a to clear outstanding I/O. Sat Sep 8 11:41:17 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
bus 10b to clear outstanding I/O.
-- --tmac
RedHat Certified Engineer #804006984323821
I didn't think the NetApps were capable of greater than 64k blocks?
Been wondering this, as I was hoping to tune up NDMP for my FC attached LTO3 drives that can handle a 256k block.
Paul
The NDMP client software specifies the size of the "mover record" -- which is the tape block size on the tape, since NetApp operates in variable block size mode. See NDMP_MOVER_SET_RECORD_SIZE. If that number is kept at dump's default of 63K bytes, then things should be fine.
(As to "use a different cable", one must use a cable rated for the higher speeds of Ultra-320, and more tightly spec'ed terminator, too. If one came with the new SCSI drive, use it.)
Steve
-----Original Message----- From: tmac [mailto:tmacmd@gmail.com] Sent: Wednesday, September 19, 2007 9:04 AM To: Ernesto Silvera Cc: toasters@mathworks.com Subject: Re: SCSI LTO3 Drive for NDMP Backup problems
From the bug ID you supplied:
Check vendor documentation for SCSI speed capabilities. If the device is capable of Ultra-320 speeds, restrict NDMP and dump/restore record sizes to 64KB or smaller when the drives are attached via the X2027A SCSI card. There is no workaround for snapmirror.
How are you doing dumps? If via NDMP you need to talk with your software vendor to figure how to change the buffer size down a little (like to 32K or something).
--tmac
On 9/19/07, Ernesto Silvera esilvera@csrven.com wrote:
We already try another cable, drive and terminator...
Netapp replies that we are hitting the Bug 153964, "SCSI transfers greater than 64KB hang on X2027A in Ultra320 mode". So the question is
what is the blocksize Netapp uses for the dump?. It is 64KB?
Regards!
-----Original Message----- From: tmac [mailto:tmacmd@gmail.com] Sent: Wednesday, September 19, 2007 8:17 AM To: Ernesto Silvera Cc: toasters@mathworks.com Subject: Re: SCSI LTO3 Drive for NDMP Backup problems
Check the SCSI connections to:
- make sure ther are no little bent pins
- screwed down TIGHT
If problems persist, try a different cable. Problems still, RMA the SCSI card. Lastly, RMA the Tape drive itself.
--tmac
On 9/19/07, Ernesto Silvera esilvera@csrven.com wrote:
Hi all,
We are getting errors while trying to replace with a new HP LTO3 drive,
the
actual LTO1 drive connected to the scsi adapter X2027A of a FAS920 NetApp Release 7.0.4 (see errors and actual configuration below).
NetApp Log # 2475693
I would really appreciate any hints..
Actual Configuration
Model Name: FAS920 NetApp Release 7.0.4
LTO1 drive connected to scsi Adapter
slot 10: SCSI Host Adapter 10a (Adaptec AIC7902, <ONLINE>)
HW Version 0.0 Ultra320 SCSI, Low Voltage Differential I/O base 0xdd00, size 0x100 memory mapped I/O base 0xa04c0000, size0x2000
I/O base 0xdf00, size 0x100 slot 10: SCSI Host Adapter 10b (Adaptec AIC7902, <ONLINE>) HW Version 0.0 Ultra320 SCSI, Low Voltage Differential 4: Tape: Seagate ULTRIUM06242 1613 I/O base 0xde00, size 0x100 memory mapped I/O base 0xa04c2000, size0x2000
I/O base 0xdc00, size 0x100SCSI ERRORS
Sat Sep 8 11:04:38 VET [adpch_timeoutd:warning]: adpch_timeoutd: 10a.3(0x0568ede0, 0x12): command timeout, quiescing bus to allow
outstanding
I/O to complete(ticks = 8). Sat Sep 8 11:04:46 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
10a target 3 to clear outstanding I/O. Sat Sep 8 11:04:46 VET [adpch_intrd:warning]: adpch_timeoutd: 10a: unquiescing bus Sat Sep 8 11:04:57 VET [adpch_timeoutd:warning]: adpch_timeoutd: 10a.3(0x0568ede0, 0x12): command timeout, quiescing bus to allow
outstanding
I/O to complete(ticks = 8). Sat Sep 8 11:05:05 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
10a target 3 to clear outstanding I/O. Sat Sep 8 11:05:05 VET [adpch_intrd:warning]: adpch_timeoutd: 10a: unquiescing bus Sat Sep 8 11:06:10 VET [adpch_intrd:warning]: Unexpected page code:
Ultrium
3 serial number not available. Sat Sep 8 11:34:49 VET [rc:info]: relog syslog Sat Sep 8 11:33:24 VET [adpch_asyncd:warning]: adpch asyncd: Resetting SCSI bus 10a to
clear
outstanding I/O. Sat Sep 8 11:34:56 VET [adpch_asyncd:warning]: adpch asyncd:
Resetting
SCSI
bus 10a to clear outstanding I/O. Sat Sep 8 11:36:28 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
bus 10a to clear outstanding I/O. Sat Sep 8 11:41:17 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
bus 10b to clear outstanding I/O.
-- --tmac
RedHat Certified Engineer #804006984323821
We had similar problems with a HP MSL5026 library attached to our GF920. Changed SCSI cards/cables, the drive in the tape library and eventually purchased an HP 6030 LTO3 lib to replace it. Still got the messages from both libraries until I disconnected the LTO2 from the system - haven't had an error like it since. The only part the I didn't replace was the library scsi controller card
regards John
Ernesto Silvera wrote:
We already try another cable, drive and terminator...
Netapp replies that we are hitting the Bug 153964, "SCSI transfers greater than 64KB hang on X2027A in Ultra320 mode". So the question is what is the blocksize Netapp uses for the dump?. It is 64KB?
Regards!
-----Original Message----- From: tmac [mailto:tmacmd@gmail.com] Sent: Wednesday, September 19, 2007 8:17 AM To: Ernesto Silvera Cc: toasters@mathworks.com Subject: Re: SCSI LTO3 Drive for NDMP Backup problems
Check the SCSI connections to:
- make sure ther are no little bent pins
- screwed down TIGHT
If problems persist, try a different cable. Problems still, RMA the SCSI card. Lastly, RMA the Tape drive itself.
--tmac
On 9/19/07, Ernesto Silvera esilvera@csrven.com wrote:
Hi all,
We are getting errors while trying to replace with a new HP LTO3 drive,
the
actual LTO1 drive connected to the scsi adapter X2027A of a FAS920 NetApp Release 7.0.4 (see errors and actual configuration below).
NetApp Log # 2475693
I would really appreciate any hints..
Actual Configuration
Model Name: FAS920 NetApp Release 7.0.4
LTO1 drive connected to scsi Adapter
slot 10: SCSI Host Adapter 10a (Adaptec AIC7902, <ONLINE>)
HW Version 0.0 Ultra320 SCSI, Low Voltage Differential I/O base 0xdd00, size 0x100 memory mapped I/O base 0xa04c0000, size 0x2000 I/O base 0xdf00, size 0x100 slot 10: SCSI Host Adapter 10b (Adaptec AIC7902, <ONLINE>) HW Version 0.0 Ultra320 SCSI, Low Voltage Differential 4: Tape: Seagate ULTRIUM06242 1613 I/O base 0xde00, size 0x100 memory mapped I/O base 0xa04c2000, size 0x2000 I/O base 0xdc00, size 0x100SCSI ERRORS
Sat Sep 8 11:04:38 VET [adpch_timeoutd:warning]: adpch_timeoutd: 10a.3(0x0568ede0, 0x12): command timeout, quiescing bus to allow
outstanding
I/O to complete(ticks = 8). Sat Sep 8 11:04:46 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
10a target 3 to clear outstanding I/O. Sat Sep 8 11:04:46 VET [adpch_intrd:warning]: adpch_timeoutd: 10a: unquiescing bus Sat Sep 8 11:04:57 VET [adpch_timeoutd:warning]: adpch_timeoutd: 10a.3(0x0568ede0, 0x12): command timeout, quiescing bus to allow
outstanding
I/O to complete(ticks = 8). Sat Sep 8 11:05:05 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
10a target 3 to clear outstanding I/O. Sat Sep 8 11:05:05 VET [adpch_intrd:warning]: adpch_timeoutd: 10a: unquiescing bus Sat Sep 8 11:06:10 VET [adpch_intrd:warning]: Unexpected page code:
Ultrium
3 serial number not available. Sat Sep 8 11:34:49 VET [rc:info]: relog syslog Sat Sep 8 11:33:24 VET [adpch_asyncd:warning]: adpch asyncd: Resetting SCSI bus 10a to clear outstanding I/O. Sat Sep 8 11:34:56 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
bus 10a to clear outstanding I/O. Sat Sep 8 11:36:28 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
bus 10a to clear outstanding I/O. Sat Sep 8 11:41:17 VET [adpch_asyncd:warning]: adpch asyncd: Resetting
SCSI
bus 10b to clear outstanding I/O.
-
Ernesto Silvera -
John Files -
Parsons, Steve -
Paul Keating -
tmac