Discussion:
[Iscsitarget-devel] cmnd_abort issues in wb mode
Ingo Flaschberger
2007-01-24 22:58:22 UTC
Permalink
Hi,

I have following setup:

target:
iscsi-target (svn version)
Gigenw-card Intel 82541PI, mtu 9000
1GB Ram
Amd XP 2500+
3ware raid5
linux 2.6.19.2
xfs partition for iscsi
host:
linux-iscsi 4:0.1.11-3
Gigenw-card Marvell 88E8001, mtu 9000
2GB Ram
Intel Pentium D CPU 2.80GHz (dualcpu)
openvz linux 2.6.9-023stab037
ext3 partition on iscsi

When I run bonnie at the host machine and configure write back (wb) mode
at the iscsi-target, the system shows abort messages.
Without the wb mode, everything is ok.

There are also some ip packets dropped at the target gige card, but not
very much (~1 per second at full traffic).
When I use the other network cards (100mbit network) wb mode is also ok.


target:
---snip---
execute_task_management(1212) 28a1 1 3d1d0000
cmnd_abort(1143) 1d3d 1 0 42 131072 0 0
execute_task_management(1212) 28a2 1 3e1d0000
cmnd_abort(1143) 1d3e 1 0 42 131072 0 0
execute_task_management(1212) 28a3 1 3f1d0000
cmnd_abort(1143) 1d3f 1 0 42 131072 0 0
execute_task_management(1212) 28a4 1 401d0000
cmnd_abort(1143) 1d40 1 0 42 131072 0 0
execute_task_management(1212) 28a5 1 411d0000
cmnd_abort(1143) 1d41 1 0 42 131072 0 0
execute_task_management(1212) 28a6 1 421d0000
cmnd_abort(1143) 1d42 1 0 42 131072 0 0
execute_task_management(1212) 28a7 1 431d0000
cmnd_abort(1143) 1d43 1 0 42 131072 0 0
execute_task_management(1212) 28a8 2 ffffffff
execute_task_management(1212) 92c1 1 6880000
cmnd_abort(1143) 8806 1 0 42 20480 0 0
execute_task_management(1212) 92c2 1 7880000
cmnd_abort(1143) 8807 1 0 42 131072 0 0
execute_task_management(1212) 92c3 1 8880000
cmnd_abort(1143) 8808 1 0 42 4096 0 0
---snap---

host:
---snip---
iscsi-sfnet:host10: eh_abort cmnd already done
iscsi-sfnet:host10: eh_abort cmnd already done
iscsi-sfnet:host10: eh_abort cmnd already done
iscsi-sfnet:host10: Sending ABORT TASK for task itt 7485
iscsi-sfnet:host10: Waking tx_thread to send task mgmt function itt 10401
iscsi-sfnet:host10: task mgmt itt 10401 rejected (0x1)
iscsi-sfnet:host10: Sending ABORT TASK for task itt 7486
iscsi-sfnet:host10: Waking tx_thread to send task mgmt function itt 10402
iscsi-sfnet:host10: task mgmt itt 10402 rejected (0x1)
iscsi-sfnet:host10: Sending ABORT TASK for task itt 7487
iscsi-sfnet:host10: Waking tx_thread to send task mgmt function itt 10403
iscsi-sfnet:host10: task mgmt itt 10403 rejected (0x1)
iscsi-sfnet:host10: Sending ABORT TASK for task itt 7488
iscsi-sfnet:host10: Waking tx_thread to send task mgmt function itt 10404
iscsi-sfnet:host10: task mgmt itt 10404 rejected (0x1)
iscsi-sfnet:host10: Sending ABORT TASK for task itt 7489
iscsi-sfnet:host10: Waking tx_thread to send task mgmt function itt 10405
iscsi-sfnet:host10: task mgmt itt 10405 rejected (0x1)
iscsi-sfnet:host10: Sending ABORT TASK for task itt 7490
iscsi-sfnet:host10: Waking tx_thread to send task mgmt function itt 10406
iscsi-sfnet:host10: task mgmt itt 10406 rejected (0x1)
iscsi-sfnet:host10: Sending ABORT TASK for task itt 7491
iscsi-sfnet:host10: Waking tx_thread to send task mgmt function itt 10407
iscsi-sfnet:host10: task mgmt itt 10407 rejected (0x1)
iscsi-sfnet:host10: Sending ABORT TASK SET
iscsi-sfnet:host10: Waking tx_thread to send task mgmt function itt 10408
iscsi-sfnet:host10: task mgmt itt 10408 successful
iscsi-sfnet:host10: Failing command cdb 0x2a task 7485 with return code = 0x20000
iscsi-sfnet:host10: Failing command cdb 0x2a task 7486 with return code = 0x20000
iscsi-sfnet:host10: Failing command cdb 0x2a task 7487 with return code = 0x20000
iscsi-sfnet:host10: Failing command cdb 0x2a task 7488 with return code = 0x20000
iscsi-sfnet:host10: Failing command cdb 0x2a task 7489 with return code = 0x20000
iscsi-sfnet:host10: Failing command cdb 0x2a task 7490 with return code = 0x20000
iscsi-sfnet:host10: Failing command cdb 0x2a task 7491 with return code = 0x20000
iscsi-sfnet:host10: recv_cmd - response for itt 7485, but no such task
iscsi-sfnet:host10: recv_cmd - response for itt 7486, but no such task
iscsi-sfnet:host10: recv_cmd - response for itt 7487, but no such task
iscsi-sfnet:host10: recv_cmd - response for itt 7488, but no such task
iscsi-sfnet:host10: recv_cmd - response for itt 7489, but no such task
iscsi-sfnet:host10: recv_cmd - response for itt 7490, but no such task
iscsi-sfnet:host10: recv_cmd - response for itt 7491, but no such task
---snap---


target config:
---snip---
IncomingUser sns xxxxxxxxxxxx
OutgoingUser storage xxxxxxxxxxxxxxx

Target iqn.2007-01.xxxx.s0004:vz
Alias s0004.crossip.net
# 123456789012
Lun 0 Path=/storage/iqn.2007-01.xxxx.s0004.disk,Type=fileio,IOMode=wb
IncomingUser sns xxxxxxxxxxx
OutgoingUser storage xxxxxxxxxx
HeaderDigest CRC32C
DataDigest CRC32C
MaxConnections 1
InitialR2T No
ImmediateData Yes
MaxRecvDataSegmentLength 131072
MaxXmitDataSegmentLength 131072
MaxBurstLength 16776192
FirstBurstLength 262144
MaxOutstandingR2T 100
DataPDUInOrder Yes
DataSequenceInOrder Yes
ErrorRecoveryLevel 0
Wthreads 8
QueuedCommands 32
---snap---

host config:
---snip---
DiscoveryAddress=192.168.100.1
OutgoingUsername=sns
OutgoingPassword=xxxxx
IncomingUsername=storage
IncomingPassword=xxxxxx

TargetName=iqn.2007-01.xxxx.s0004:vz
HeaderDigest=always
DataDigest=always
ImmediateData=yes
MaxRecvDataSegmentLength=131072
FirstBurstLength=262144
MaxBurstLength=16776192
AbortTimeout=60
ConnFailTimeout=0
---snap---


Any ideas where to search, screw at the config?

bye,
ingo flaschberger

geschaeftsleitung
---------------------------
netstorage-crossip-flat:fee
powered by
crossip communications gmbh
---------------------------
sebastian kneipp gasse 1
a-1020 wien
fix: +43-1-726 15 22-217
fax: +43-1-726 15 22-111
---------------------------
Ming Zhang
2007-01-25 00:26:52 UTC
Permalink
Post by Ingo Flaschberger
Hi,
iscsi-target (svn version)
Gigenw-card Intel 82541PI, mtu 9000
1GB Ram
Amd XP 2500+
3ware raid5
linux 2.6.19.2
xfs partition for iscsi
could u try a raw parition from 3ware?
Post by Ingo Flaschberger
linux-iscsi 4:0.1.11-3
Gigenw-card Marvell 88E8001, mtu 9000
2GB Ram
Intel Pentium D CPU 2.80GHz (dualcpu)
openvz linux 2.6.9-023stab037
ext3 partition on iscsi
When I run bonnie at the host machine and configure write back (wb) mode
at the iscsi-target, the system shows abort messages.
Without the wb mode, everything is ok.
There are also some ip packets dropped at the target gige card, but not
very much (~1 per second at full traffic).
When I use the other network cards (100mbit network) wb mode is also ok.
---snip---
execute_task_management(1212) 28a1 1 3d1d0000
cmnd_abort(1143) 1d3d 1 0 42 131072 0 0
execute_task_management(1212) 28a2 1 3e1d0000
cmnd_abort(1143) 1d3e 1 0 42 131072 0 0
execute_task_management(1212) 28a3 1 3f1d0000
cmnd_abort(1143) 1d3f 1 0 42 131072 0 0
execute_task_management(1212) 28a4 1 401d0000
cmnd_abort(1143) 1d40 1 0 42 131072 0 0
execute_task_management(1212) 28a5 1 411d0000
cmnd_abort(1143) 1d41 1 0 42 131072 0 0
execute_task_management(1212) 28a6 1 421d0000
cmnd_abort(1143) 1d42 1 0 42 131072 0 0
execute_task_management(1212) 28a7 1 431d0000
cmnd_abort(1143) 1d43 1 0 42 131072 0 0
execute_task_management(1212) 28a8 2 ffffffff
execute_task_management(1212) 92c1 1 6880000
cmnd_abort(1143) 8806 1 0 42 20480 0 0
execute_task_management(1212) 92c2 1 7880000
cmnd_abort(1143) 8807 1 0 42 131072 0 0
execute_task_management(1212) 92c3 1 8880000
cmnd_abort(1143) 8808 1 0 42 4096 0 0
---snap---
---snip---
iscsi-sfnet:host10: eh_abort cmnd already done
iscsi-sfnet:host10: eh_abort cmnd already done
iscsi-sfnet:host10: eh_abort cmnd already done
iscsi-sfnet:host10: Sending ABORT TASK for task itt 7485
iscsi-sfnet:host10: Waking tx_thread to send task mgmt function itt 10401
iscsi-sfnet:host10: task mgmt itt 10401 rejected (0x1)
iscsi-sfnet:host10: Sending ABORT TASK for task itt 7486
iscsi-sfnet:host10: Waking tx_thread to send task mgmt function itt 10402
iscsi-sfnet:host10: task mgmt itt 10402 rejected (0x1)
iscsi-sfnet:host10: Sending ABORT TASK for task itt 7487
iscsi-sfnet:host10: Waking tx_thread to send task mgmt function itt 10403
iscsi-sfnet:host10: task mgmt itt 10403 rejected (0x1)
iscsi-sfnet:host10: Sending ABORT TASK for task itt 7488
iscsi-sfnet:host10: Waking tx_thread to send task mgmt function itt 10404
iscsi-sfnet:host10: task mgmt itt 10404 rejected (0x1)
iscsi-sfnet:host10: Sending ABORT TASK for task itt 7489
iscsi-sfnet:host10: Waking tx_thread to send task mgmt function itt 10405
iscsi-sfnet:host10: task mgmt itt 10405 rejected (0x1)
iscsi-sfnet:host10: Sending ABORT TASK for task itt 7490
iscsi-sfnet:host10: Waking tx_thread to send task mgmt function itt 10406
iscsi-sfnet:host10: task mgmt itt 10406 rejected (0x1)
iscsi-sfnet:host10: Sending ABORT TASK for task itt 7491
iscsi-sfnet:host10: Waking tx_thread to send task mgmt function itt 10407
iscsi-sfnet:host10: task mgmt itt 10407 rejected (0x1)
iscsi-sfnet:host10: Sending ABORT TASK SET
iscsi-sfnet:host10: Waking tx_thread to send task mgmt function itt 10408
iscsi-sfnet:host10: task mgmt itt 10408 successful
iscsi-sfnet:host10: Failing command cdb 0x2a task 7485 with return code = 0x20000
iscsi-sfnet:host10: Failing command cdb 0x2a task 7486 with return code = 0x20000
iscsi-sfnet:host10: Failing command cdb 0x2a task 7487 with return code = 0x20000
iscsi-sfnet:host10: Failing command cdb 0x2a task 7488 with return code = 0x20000
iscsi-sfnet:host10: Failing command cdb 0x2a task 7489 with return code = 0x20000
iscsi-sfnet:host10: Failing command cdb 0x2a task 7490 with return code = 0x20000
iscsi-sfnet:host10: Failing command cdb 0x2a task 7491 with return code = 0x20000
iscsi-sfnet:host10: recv_cmd - response for itt 7485, but no such task
iscsi-sfnet:host10: recv_cmd - response for itt 7486, but no such task
iscsi-sfnet:host10: recv_cmd - response for itt 7487, but no such task
iscsi-sfnet:host10: recv_cmd - response for itt 7488, but no such task
iscsi-sfnet:host10: recv_cmd - response for itt 7489, but no such task
iscsi-sfnet:host10: recv_cmd - response for itt 7490, but no such task
iscsi-sfnet:host10: recv_cmd - response for itt 7491, but no such task
---snap---
---snip---
IncomingUser sns xxxxxxxxxxxx
OutgoingUser storage xxxxxxxxxxxxxxx
Target iqn.2007-01.xxxx.s0004:vz
Alias s0004.crossip.net
# 123456789012
Lun 0 Path=/storage/iqn.2007-01.xxxx.s0004.disk,Type=fileio,IOMode=wb
IncomingUser sns xxxxxxxxxxx
OutgoingUser storage xxxxxxxxxx
HeaderDigest CRC32C
DataDigest CRC32C
MaxConnections 1
InitialR2T No
ImmediateData Yes
MaxRecvDataSegmentLength 131072
MaxXmitDataSegmentLength 131072
MaxBurstLength 16776192
FirstBurstLength 262144
MaxOutstandingR2T 100
DataPDUInOrder Yes
DataSequenceInOrder Yes
ErrorRecoveryLevel 0
Wthreads 8
QueuedCommands 32
---snap---
---snip---
DiscoveryAddress=192.168.100.1
OutgoingUsername=sns
OutgoingPassword=xxxxx
IncomingUsername=storage
IncomingPassword=xxxxxx
TargetName=iqn.2007-01.xxxx.s0004:vz
HeaderDigest=always
DataDigest=always
ImmediateData=yes
MaxRecvDataSegmentLength=131072
FirstBurstLength=262144
MaxBurstLength=16776192
AbortTimeout=60
ConnFailTimeout=0
---snap---
Any ideas where to search, screw at the config?
bye,
ingo flaschberger
geschaeftsleitung
---------------------------
netstorage-crossip-flat:fee
powered by
crossip communications gmbh
---------------------------
sebastian kneipp gasse 1
a-1020 wien
fix: +43-1-726 15 22-217
fax: +43-1-726 15 22-111
---------------------------
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Iscsitarget-devel mailing list
https://lists.sourceforge.net/lists/listinfo/iscsitarget-devel
--
http://blackmagic02881.wordpress.com/
Ingo Flaschberger
2007-01-25 18:53:01 UTC
Permalink
Hi,
Post by Ming Zhang
Post by Ingo Flaschberger
iscsi-target (svn version)
Gigenw-card Intel 82541PI, mtu 9000
1GB Ram
Amd XP 2500+
3ware raid5
linux 2.6.19.2
xfs partition for iscsi
could u try a raw parition from 3ware?
Yes, same errors.
The iscsi host starts first with the eh_abort, and later the iscsi-target
reacts at the aborts.

Without the wb-mode, everything is fine.
I have also tried to set the AbortTimeout=120 very high, no success.

bye,
Ingo
Ming Zhang
2007-01-25 18:56:00 UTC
Permalink
Post by Ingo Flaschberger
Hi,
Post by Ming Zhang
Post by Ingo Flaschberger
iscsi-target (svn version)
Gigenw-card Intel 82541PI, mtu 9000
1GB Ram
Amd XP 2500+
3ware raid5
linux 2.6.19.2
could u try to step down kernel version to .16 or 18 and see if still
can reproduce? i do not have HW RAID controller and my MD so far perform
ok.

and i do not see people report this before.
Post by Ingo Flaschberger
Post by Ming Zhang
Post by Ingo Flaschberger
xfs partition for iscsi
could u try a raw parition from 3ware?
Yes, same errors.
The iscsi host starts first with the eh_abort, and later the iscsi-target
reacts at the aborts.
Without the wb-mode, everything is fine.
I have also tried to set the AbortTimeout=120 very high, no success.
bye,
Ingo
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Iscsitarget-devel mailing list
https://lists.sourceforge.net/lists/listinfo/iscsitarget-devel
--
http://blackmagic02881.wordpress.com/
Ingo Flaschberger
2007-01-25 20:43:41 UTC
Permalink
Hi,
Post by Ming Zhang
Post by Ingo Flaschberger
iscsi-target (svn version)
Gigenw-card Intel 82541PI, mtu 9000
1GB Ram
Amd XP 2500+
3ware raid5
linux 2.6.19.2
could u try to step down kernel version to .16 or 18 and see if still
can reproduce? i do not have HW RAID controller and my MD so far perform
ok.
Tried vanilla 2.6.16 and no errors occur.

bye,
Ingo
Ming Zhang
2007-01-25 20:59:04 UTC
Permalink
Post by Ingo Flaschberger
Hi,
Post by Ming Zhang
Post by Ingo Flaschberger
iscsi-target (svn version)
Gigenw-card Intel 82541PI, mtu 9000
1GB Ram
Amd XP 2500+
3ware raid5
linux 2.6.19.2
could u try to step down kernel version to .16 or 18 and see if still
can reproduce? i do not have HW RAID controller and my MD so far perform
ok.
Tried vanilla 2.6.16 and no errors occur.
could you try 2.6.18 again, closer to 2.6.19.2. and try to keep kernel
configuration as close as possible. thx.
Post by Ingo Flaschberger
bye,
Ingo
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Iscsitarget-devel mailing list
https://lists.sourceforge.net/lists/listinfo/iscsitarget-devel
--
http://blackmagic02881.wordpress.com/
Ingo Flaschberger
2007-01-25 23:41:53 UTC
Permalink
Hi,
Post by Ming Zhang
Post by Ingo Flaschberger
Post by Ming Zhang
Post by Ingo Flaschberger
iscsi-target (svn version)
Gigenw-card Intel 82541PI, mtu 9000
1GB Ram
Amd XP 2500+
3ware raid5
linux 2.6.19.2
could u try to step down kernel version to .16 or 18 and see if still
can reproduce? i do not have HW RAID controller and my MD so far perform
ok.
Tried vanilla 2.6.16 and no errors occur.
could you try 2.6.18 again, closer to 2.6.19.2. and try to keep kernel
configuration as close as possible. thx.
Tried vanilla 2.6.18 and no errors.
Used make oldconfig.

bye,
Ingo
Ming Zhang
2007-01-26 00:08:15 UTC
Permalink
Post by Ingo Flaschberger
Hi,
Post by Ming Zhang
Post by Ingo Flaschberger
Post by Ming Zhang
Post by Ingo Flaschberger
iscsi-target (svn version)
Gigenw-card Intel 82541PI, mtu 9000
1GB Ram
Amd XP 2500+
3ware raid5
linux 2.6.19.2
could u try to step down kernel version to .16 or 18 and see if still
can reproduce? i do not have HW RAID controller and my MD so far perform
ok.
Tried vanilla 2.6.16 and no errors occur.
could you try 2.6.18 again, closer to 2.6.19.2. and try to keep kernel
configuration as close as possible. thx.
Tried vanilla 2.6.18 and no errors.
Used make oldconfig.
ic. so it probably because the change to fit new kernel api change from
generic_file_read/write to that do_sync_read/write.

please stick to pre 2.6.19 kernel till we get a solution. thx
Post by Ingo Flaschberger
bye,
Ingo
--
http://blackmagic02881.wordpress.com/
Ingo Flaschberger
2007-01-26 00:44:55 UTC
Permalink
Post by Ming Zhang
Post by Ingo Flaschberger
Post by Ming Zhang
Post by Ingo Flaschberger
Tried vanilla 2.6.16 and no errors occur.
could you try 2.6.18 again, closer to 2.6.19.2. and try to keep kernel
configuration as close as possible. thx.
Tried vanilla 2.6.18 and no errors.
Used make oldconfig.
ic. so it probably because the change to fit new kernel api change from
generic_file_read/write to that do_sync_read/write.
please stick to pre 2.6.19 kernel till we get a solution. thx
Ok.
Anything else I can do? (Read and write C)

bye,
Ingo
Ming Zhang
2007-01-26 14:06:28 UTC
Permalink
Post by Ingo Flaschberger
Post by Ming Zhang
Post by Ingo Flaschberger
Post by Ming Zhang
Post by Ingo Flaschberger
Tried vanilla 2.6.16 and no errors occur.
could you try 2.6.18 again, closer to 2.6.19.2. and try to keep kernel
configuration as close as possible. thx.
Tried vanilla 2.6.18 and no errors.
Used make oldconfig.
ic. so it probably because the change to fit new kernel api change from
generic_file_read/write to that do_sync_read/write.
please stick to pre 2.6.19 kernel till we get a solution. thx
Ok.
Anything else I can do? (Read and write C)
i have to reproduce this but need some time. or maybe someone can help
you quicker.
Post by Ingo Flaschberger
bye,
Ingo
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Iscsitarget-devel mailing list
https://lists.sourceforge.net/lists/listinfo/iscsitarget-devel
--
http://blackmagic02881.wordpress.com/
Ming Zhang
2007-01-29 00:41:25 UTC
Permalink
Post by Ming Zhang
Post by Ingo Flaschberger
Post by Ming Zhang
Post by Ingo Flaschberger
Post by Ming Zhang
Post by Ingo Flaschberger
Tried vanilla 2.6.16 and no errors occur.
could you try 2.6.18 again, closer to 2.6.19.2. and try to keep kernel
configuration as close as possible. thx.
Tried vanilla 2.6.18 and no errors.
Used make oldconfig.
ic. so it probably because the change to fit new kernel api change from
generic_file_read/write to that do_sync_read/write.
please stick to pre 2.6.19 kernel till we get a solution. thx
Ok.
Anything else I can do? (Read and write C)
i have to reproduce this but need some time. or maybe someone can help
you quicker.
btw, if this is the real problem spot. it will be great if we jump to
change the current read/write to readv/writev directly.
Post by Ming Zhang
Post by Ingo Flaschberger
bye,
Ingo
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Iscsitarget-devel mailing list
https://lists.sourceforge.net/lists/listinfo/iscsitarget-devel
--
http://blackmagic02881.wordpress.com/
Ingo Flaschberger
2007-01-25 23:45:23 UTC
Permalink
Hi

..and 2.6.18 is much faster:
2.6.18 iscsi to raw disk, wb mode, no errors
Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
s0004.crossip.ne 4G 25522 62 21005 9 9262 2 19711 38 29521 3 121.9 0
------Sequential Create------ --------Random Create--------
-Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
16 2600 98 +++++ +++ +++++ +++ 2557 96 +++++ +++ 8331 99
s0004.crossip.net,4G,25522,62,21005,9,9262,2,19711,38,29521,3,121.9,0,16,2600,98,+++++,+++,+++++,+++,2557,96,+++++,+++,8331,99

2.6.19.2 iscsi to raw disk, wb mode, errors
Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
s0004.crossip.ne 4G 17562 42 16835 7 8833 2 19124 37 29387 3 121.8 0
------Sequential Create------ --------Random Create--------
-Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
16 2605 98 +++++ +++ +++++ +++ 2603 97 +++++ +++ 8212 100
s0004.crossip.net,4G,17562,42,16835,7,8833,2,19124,37,29387,3,121.8,0,16,2605,98,+++++,+++,+++++,+++,2603,97,+++++,+++,8212,100

bye,
Ingo
Ming Zhang
2007-01-26 00:09:35 UTC
Permalink
Post by Ingo Flaschberger
Hi
2.6.18 iscsi to raw disk, wb mode, no errors
Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
s0004.crossip.ne 4G 25522 62 21005 9 9262 2 19711 38 29521 3 121.9 0
------Sequential Create------ --------Random Create--------
-Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
16 2600 98 +++++ +++ +++++ +++ 2557 96 +++++ +++ 8331 99
s0004.crossip.net,4G,25522,62,21005,9,9262,2,19711,38,29521,3,121.9,0,16,2600,98,+++++,+++,+++++,+++,2557,96,+++++,+++,8331,99
even without error, your results are not fast at all.
Post by Ingo Flaschberger
2.6.19.2 iscsi to raw disk, wb mode, errors
Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
s0004.crossip.ne 4G 17562 42 16835 7 8833 2 19124 37 29387 3 121.8 0
------Sequential Create------ --------Random Create--------
-Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
16 2605 98 +++++ +++ +++++ +++ 2603 97 +++++ +++ 8212 100
s0004.crossip.net,4G,17562,42,16835,7,8833,2,19124,37,29387,3,121.8,0,16,2605,98,+++++,+++,+++++,+++,2603,97,+++++,+++,8212,100
bye,
Ingo
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Iscsitarget-devel mailing list
https://lists.sourceforge.net/lists/listinfo/iscsitarget-devel
--
http://blackmagic02881.wordpress.com/
Ingo Flaschberger
2007-01-26 00:47:23 UTC
Permalink
Hi,
Post by Ming Zhang
Post by Ingo Flaschberger
2.6.18 iscsi to raw disk, wb mode, no errors
Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP
s0004.crossip.ne 4G 25522 62 21005 9 9262 2 19711 38 29521 3 121.9 0
------Sequential Create------ --------Random Create--------
-Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
16 2600 98 +++++ +++ +++++ +++ 2557 96 +++++ +++ 8331 99
s0004.crossip.net,4G,25522,62,21005,9,9262,2,19711,38,29521,3,121.9,0,16,2600,98,+++++,+++,+++++,+++,2557,96,+++++,+++,8331,99
even without error, your results are not fast at all.
I know, the 8005 3ware controllers are a little bit "lame".
bonnie at the target machine directly is not much faster.

bye,
Ingo
Loading...