Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Fusion MPT LSI53C1030 RAID very slow
View unanswered posts
View posts from last 24 hours

 
Reply to topic    Gentoo Forums Forum Index Gentoo on AMD64
View previous topic :: View next topic  
Author Message
zeek
Guru
Guru


Joined: 16 Nov 2002
Posts: 480
Location: Bantayan Island

PostPosted: Wed Jan 12, 2005 1:39 am    Post subject: Fusion MPT LSI53C1030 RAID very slow Reply with quote

I've got 2 new servers:

IBM eServer 326 (1U)
Opteron 2.4 GHz
5 GB RAM
2 x IBM 146.8 GB 10K RPM Ultra320 SCSI Hot-Swap HDD (mirrored)

http://www-1.ibm.com/servers/eserver/opteron/325/326_more_info.html

and the disk performance is very slow. In fact it seems to be more than 20x times slower than a comparable P4 using a different RAID controller.

IBM LSI53C1030 RAID 1:
Code:
tux mysql-super-smack # dbench -c /usr/share/dbench/client_oplocks.txt 100
100 clients started
   0     62477  10.50 MB/seccc
Throughput 10.4985 MB/sec 100 procs


DELL 1850 PERC 4 RAID 1 (LSI megaraid):
Code:
tux mysql-super-smack # dbench -c /usr/share/dbench/client_oplocks.txt 100
100 clients started
   0     62477  225.86 MB/sec
Throughput 225.861 MB/sec 100 procs


hdparm looks normal, hdparm -Tt on the IBM LSI53C1030 RAID 1:
Code:
tux lsi # hdparm -tT /dev/sda

/dev/sda:
 Timing cached reads:   3600 MB in  2.00 seconds = 1800.27 MB/sec
 Timing buffered disk reads:  192 MB in  3.02 seconds =  63.67 MB/sec


And some bad news from google:

http://www.ussg.iu.edu/hypermail/linux/kernel/0311.3/0131.html
http://seclists.org/lists/linux-kernel/2004/Jul/5767.html

I have the iommu=merge option on the kernel line as directed in the Gentoo AMD64 technotes.

If you have one of these Fusion MPT controllers, can you emerge dbench are run it with the options I have above and post the results please? I'd like to see if everyone who has one of these controllers is suffering poor disk performance.

Thanks!
Back to top
View user's profile Send private message
[dmnd]
n00b
n00b


Joined: 02 Nov 2003
Posts: 48
Location: Netherlands

PostPosted: Wed Jan 12, 2005 7:10 am    Post subject: Reply with quote

Code:
ns0(root) ~ # cat /proc/cpuinfo
processor       : 0
vendor_id       : GenuineIntel
cpu family      : 15
model           : 2
model name      : Intel(R) Xeon(TM) CPU 2.40GHz
stepping        : 9
cpu MHz         : 2392.567
cache size      : 512 KB
fdiv_bug        : no
hlt_bug         : no
f00f_bug        : no
coma_bug        : no
fpu             : yes
fpu_exception   : yes
cpuid level     : 2
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm
bogomips        : 4771.02

ns0(root) ~ #



Code:
ns0(root) ~ # hdparm -tT /dev/sda

/dev/sda:
 Timing buffer-cache reads:   1920 MB in  2.00 seconds = 960.00 MB/sec
 Timing buffered disk reads:  218 MB in  3.00 seconds =  72.55 MB/sec
ns0(root) ~ #



Code:
ns0(root) ~ # dbench -c /usr/share/dbench/client_oplocks.txt 15
15 clients started
   0     62477  66.29 MB/secc
Throughput 66.2901 MB/sec 15 procs
ns0(root) ~ #


Also, please check your inbox :)
_________________
cold as ice...
Back to top
View user's profile Send private message
zeek
Guru
Guru


Joined: 16 Nov 2002
Posts: 480
Location: Bantayan Island

PostPosted: Wed Jan 12, 2005 7:21 pm    Post subject: Reply with quote

[dmnd] wrote:
Code:
ns0(root) ~ # dbench -c /usr/share/dbench/client_oplocks.txt 15
15 clients started
   0     62477  66.29 MB/secc
Throughput 66.2901 MB/sec 15 procs
ns0(root) ~ #



The value you get seems excessivly low. My Fusion MPT gets:
Code:
tux freeipmi # dbench -c /usr/share/dbench/client_oplocks.txt 15
15 clients started
   0     62477  788.75 MB/secc
Throughput 788.74 MB/sec 15 procs


and the Dell Megaraid:
Code:
tux freeipmi # dbench -c /usr/share/dbench/client_oplocks.txt 15
15 clients started
   0     62477  1063.49 MB/sec
Throughput 1063.48 MB/sec 15 procs


It seems that the higher the clients go, the more the performance drops off as compared to different controllers.
Back to top
View user's profile Send private message
gEistiO
n00b
n00b


Joined: 17 Mar 2005
Posts: 5

PostPosted: Thu Mar 17, 2005 7:31 pm    Post subject: Reply with quote

Hi!

I know that this is just another ancient thread, but as I've got exactly the same problem, I wondered if anyone already found a solution...
To be precise, I own a Dell PowerEdge 1600SC with the integrated PCI-X SCSI-controller LSI53C1030.
For now I just got a single harddisk, the Seagate ST336607LW.
Maybe I should also give you some information about the software:
Code:

wubbla@meister:~$ uname -a
Linux meister 2.6.11.2-poweredge #6 SMP Thu Mar 10 16:38:04 CET 2005 i686 unknown

wubbla@meister:~$ cat /proc/mpt/ioc0/*
ioc0:  (f/w download boot flag set)
  ProductID = 0x0202 (LSI53C1030)
  FWVersion = 0x01032300 (fw_size=40048)
  MsgVersion = 0x0102
  FirstWhoInit = 0x00
  EventState = 0x01
  CurrentHostMfaHighAddr = 0x00000000
  CurrentSenseBufferHighAddr = 0x00000000
  MaxChainDepth = 0x30 frames
  MinBlockSize = 0x20 bytes
  RequestFrames @ 0xdfd22000 (Dma @ 0x1fd22000)
    {CurReqSz=96} x {CurReqDepth=255} = 24480 bytes ^= 0x7000
    {MaxReqSz=96}   {MaxReqDepth=255}
  Frames   @ 0xdfd20000 (Dma @ 0x1fd20000)
    {CurRepSz=64} x {CurRepDepth=128} = 8192 bytes ^= 0x2080
    {MaxRepSz=128}   {MaxRepDepth=255}
  MaxDevices = 16
  MaxBuses = 1
  PortNumber = 1 (of 1)
ioc0: LSI53C1030, FwRev=01032300h, Ports=1, MaxQ=255, IRQ=29

wubbla@meister:~$ cat /proc/scsi/scsi
Attached devices:
Host: scsi0 Channel: 00 Id: 00 Lun: 00
  Vendor: SEAGATE  Model: ST336607LW       Rev: DS08
  Type:   Direct-Access                    ANSI SCSI revision: 03


If needed, I could also post my .config...
However, my problem now is the very poor performance of the system:
Code:

wubbla@meister:~$ sudo hdparm -tT /dev/sda

/dev/sda:
 Timing buffer-cache reads:   1832 MB in  2.00 seconds = 915.22 MB/sec
 Timing buffered disk reads:   94 MB in  3.03 seconds =  31.02 MB/sec
Back to top
View user's profile Send private message
[dmnd]
n00b
n00b


Joined: 02 Nov 2003
Posts: 48
Location: Netherlands

PostPosted: Sat Mar 19, 2005 11:41 am    Post subject: Reply with quote

gEistiO wrote:
Hi!

I know that this is just another ancient thread, but as I've got exactly the same problem, I wondered if anyone already found a solution...
To be precise, I own a Dell PowerEdge 1600SC with the integrated PCI-X SCSI-controller LSI53C1030.
For now I just got a single harddisk, the Seagate ST336607LW.
Maybe I should also give you some information about the software:
Code:

wubbla@meister:~$ uname -a
Linux meister 2.6.11.2-poweredge #6 SMP Thu Mar 10 16:38:04 CET 2005 i686 unknown

wubbla@meister:~$ cat /proc/mpt/ioc0/*
ioc0:  (f/w download boot flag set)
  ProductID = 0x0202 (LSI53C1030)
  FWVersion = 0x01032300 (fw_size=40048)
  MsgVersion = 0x0102
  FirstWhoInit = 0x00
  EventState = 0x01
  CurrentHostMfaHighAddr = 0x00000000
  CurrentSenseBufferHighAddr = 0x00000000
  MaxChainDepth = 0x30 frames
  MinBlockSize = 0x20 bytes
  RequestFrames @ 0xdfd22000 (Dma @ 0x1fd22000)
    {CurReqSz=96} x {CurReqDepth=255} = 24480 bytes ^= 0x7000
    {MaxReqSz=96}   {MaxReqDepth=255}
  Frames   @ 0xdfd20000 (Dma @ 0x1fd20000)
    {CurRepSz=64} x {CurRepDepth=128} = 8192 bytes ^= 0x2080
    {MaxRepSz=128}   {MaxRepDepth=255}
  MaxDevices = 16
  MaxBuses = 1
  PortNumber = 1 (of 1)
ioc0: LSI53C1030, FwRev=01032300h, Ports=1, MaxQ=255, IRQ=29

wubbla@meister:~$ cat /proc/scsi/scsi
Attached devices:
Host: scsi0 Channel: 00 Id: 00 Lun: 00
  Vendor: SEAGATE  Model: ST336607LW       Rev: DS08
  Type:   Direct-Access                    ANSI SCSI revision: 03


If needed, I could also post my .config...
However, my problem now is the very poor performance of the system:
Code:

wubbla@meister:~$ sudo hdparm -tT /dev/sda

/dev/sda:
 Timing buffer-cache reads:   1832 MB in  2.00 seconds = 915.22 MB/sec
 Timing buffered disk reads:   94 MB in  3.03 seconds =  31.02 MB/sec


Want to have my .config?
_________________
cold as ice...
Back to top
View user's profile Send private message
[dmnd]
n00b
n00b


Joined: 02 Nov 2003
Posts: 48
Location: Netherlands

PostPosted: Sat Mar 19, 2005 11:43 am    Post subject: Reply with quote

gEistiO wrote:
Hi!

I know that this is just another ancient thread, but as I've got exactly the same problem, I wondered if anyone already found a solution...
To be precise, I own a Dell PowerEdge 1600SC with the integrated PCI-X SCSI-controller LSI53C1030.
For now I just got a single harddisk, the Seagate ST336607LW.
Maybe I should also give you some information about the software:
Code:

wubbla@meister:~$ uname -a
Linux meister 2.6.11.2-poweredge #6 SMP Thu Mar 10 16:38:04 CET 2005 i686 unknown

wubbla@meister:~$ cat /proc/mpt/ioc0/*
ioc0:  (f/w download boot flag set)
  ProductID = 0x0202 (LSI53C1030)
  FWVersion = 0x01032300 (fw_size=40048)
  MsgVersion = 0x0102
  FirstWhoInit = 0x00
  EventState = 0x01
  CurrentHostMfaHighAddr = 0x00000000
  CurrentSenseBufferHighAddr = 0x00000000
  MaxChainDepth = 0x30 frames
  MinBlockSize = 0x20 bytes
  RequestFrames @ 0xdfd22000 (Dma @ 0x1fd22000)
    {CurReqSz=96} x {CurReqDepth=255} = 24480 bytes ^= 0x7000
    {MaxReqSz=96}   {MaxReqDepth=255}
  Frames   @ 0xdfd20000 (Dma @ 0x1fd20000)
    {CurRepSz=64} x {CurRepDepth=128} = 8192 bytes ^= 0x2080
    {MaxRepSz=128}   {MaxRepDepth=255}
  MaxDevices = 16
  MaxBuses = 1
  PortNumber = 1 (of 1)
ioc0: LSI53C1030, FwRev=01032300h, Ports=1, MaxQ=255, IRQ=29

wubbla@meister:~$ cat /proc/scsi/scsi
Attached devices:
Host: scsi0 Channel: 00 Id: 00 Lun: 00
  Vendor: SEAGATE  Model: ST336607LW       Rev: DS08
  Type:   Direct-Access                    ANSI SCSI revision: 03


If needed, I could also post my .config...
However, my problem now is the very poor performance of the system:
Code:

wubbla@meister:~$ sudo hdparm -tT /dev/sda

/dev/sda:
 Timing buffer-cache reads:   1832 MB in  2.00 seconds = 915.22 MB/sec
 Timing buffered disk reads:   94 MB in  3.03 seconds =  31.02 MB/sec


Want to have my .config?
_________________
cold as ice...
Back to top
View user's profile Send private message
gEistiO
n00b
n00b


Joined: 17 Mar 2005
Posts: 5

PostPosted: Sat Mar 19, 2005 12:04 pm    Post subject: Reply with quote

[dmnd]:

Yeah! Of course I want it! 8) :D
Back to top
View user's profile Send private message
gEistiO
n00b
n00b


Joined: 17 Mar 2005
Posts: 5

PostPosted: Wed Mar 23, 2005 4:14 pm    Post subject: Reply with quote

By the way, I've found a solution for this issue...
All you have to do is get the latest prepatch for the stable kernel, namely 2.6.12-rc1, and apply the following patch:

Code:

diff -uarN drivers/message/fusion-3.01.19/mptbase.h drivers/message/fusion/mptbase.h
--- drivers/message/fusion-3.01.19/mptbase.h   2005-03-22 10:43:07.000000000 -0700
+++ drivers/message/fusion/mptbase.h   2005-03-22 10:44:24.000000000 -0700
@@ -83,8 +83,8 @@
 #define COPYRIGHT   "Copyright (c) 1999-2004 " MODULEAUTHOR
 #endif
 
-#define MPT_LINUX_VERSION_COMMON   "3.01.19"
-#define MPT_LINUX_PACKAGE_NAME      "@(#)mptlinux-3.01.19"
+#define MPT_LINUX_VERSION_COMMON   "3.01.20"
+#define MPT_LINUX_PACKAGE_NAME      "@(#)mptlinux-3.01.20"
 #define WHAT_MAGIC_STRING      "@" "(" "#" ")"
 
 #define show_mptmod_ver(s,ver)  \
diff -uarN drivers/message/fusion-3.01.19/mptscsih.c drivers/message/fusion/mptscsih.c
--- drivers/message/fusion-3.01.19/mptscsih.c   2005-03-22 10:43:07.000000000 -0700
+++ drivers/message/fusion/mptscsih.c   2005-03-22 10:51:55.000000000 -0700
@@ -96,23 +96,26 @@
 MODULE_DESCRIPTION(my_NAME);
 MODULE_LICENSE("GPL");
 
-#ifdef MODULE
-static int dv = MPTSCSIH_DOMAIN_VALIDATION;
-module_param(dv, int, 0);
-MODULE_PARM_DESC(dv, "DV Algorithm: enhanced = 1, basic = 0 (default=MPTSCSIH_DOMAIN_VALIDATION=1)");
-
-static int width = MPTSCSIH_MAX_WIDTH;
-module_param(width, int, 0);
-MODULE_PARM_DESC(width, "Max Bus Width: wide = 1, narrow = 0 (default=MPTSCSIH_MAX_WIDTH=1)");
-
-static ushort factor = MPTSCSIH_MIN_SYNC;
-module_param(factor, ushort, 0);
-MODULE_PARM_DESC(factor, "Min Sync Factor: (default=MPTSCSIH_MIN_SYNC=0x08)");
-
-static int saf_te = MPTSCSIH_SAF_TE;
-module_param(saf_te, int, 0);
-MODULE_PARM_DESC(saf_te, "Force enabling SEP Processor: (default=MPTSCSIH_SAF_TE=0)");
-#endif
+/* Command line args */
+static int mpt_dv = MPTSCSIH_DOMAIN_VALIDATION;
+MODULE_PARM(mpt_dv, "i");
+MODULE_PARM_DESC(mpt_dv, " DV Algorithm: enhanced=1, basic=0 (default=MPTSCSIH_DOMAIN_VALIDATION=1)");
+
+static int mpt_width = MPTSCSIH_MAX_WIDTH;
+MODULE_PARM(mpt_width, "i");
+MODULE_PARM_DESC(mpt_width, " Max Bus Width: wide=1, narrow=0 (default=MPTSCSIH_MAX_WIDTH=1)");
+
+static int mpt_factor = MPTSCSIH_MIN_SYNC;
+MODULE_PARM(mpt_factor, "h");
+MODULE_PARM_DESC(mpt_factor, " Min Sync Factor (default=MPTSCSIH_MIN_SYNC=0x08)");
+
+static int mpt_saf_te = MPTSCSIH_SAF_TE;
+MODULE_PARM(mpt_saf_te, "i");
+MODULE_PARM_DESC(mpt_saf_te, " Force enabling SEP Processor: enable=1  (default=MPTSCSIH_SAF_TE=0)");
+
+static int mpt_pq_filter = 0;
+MODULE_PARM(mpt_pq_filter, "i");
+MODULE_PARM_DESC(mpt_pq_filter, " Enable peripheral qualifier filter: enable=1  (default=0)");
 
 /*=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=*/
 
@@ -253,7 +256,6 @@
 
 /* Driver command line structure
  */
-static struct mptscsih_driver_setup driver_setup;
 static struct scsi_host_template driver_template;
 
 /*=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=*/
@@ -1224,29 +1226,23 @@
       /* Update with the driver setup
        * values.
        */
-      if (ioc->spi_data.maxBusWidth >
-        driver_setup.max_width) {
-         ioc->spi_data.maxBusWidth =
-           driver_setup.max_width;
-      }
-
-      if (ioc->spi_data.minSyncFactor <
-        driver_setup.min_sync_factor) {
-         ioc->spi_data.minSyncFactor =
-           driver_setup.min_sync_factor;
-      }
+      if (ioc->spi_data.maxBusWidth > mpt_width)
+         ioc->spi_data.maxBusWidth = mpt_width;
+      if (ioc->spi_data.minSyncFactor < mpt_factor)
+         ioc->spi_data.minSyncFactor = mpt_factor;
 
       if (ioc->spi_data.minSyncFactor == MPT_ASYNC) {
          ioc->spi_data.maxSyncOffset = 0;
       }
 
-      ioc->spi_data.Saf_Te = driver_setup.saf_te;
+      ioc->spi_data.Saf_Te = mpt_saf_te;
 
       hd->negoNvram = 0;
 #ifndef MPTSCSIH_ENABLE_DOMAIN_VALIDATION
       hd->negoNvram = MPT_SCSICFG_USE_NVRAM;
 #endif
       ioc->spi_data.forceDv = 0;
+      ioc->spi_data.noQas = 0;
       for (ii=0; ii < MPT_MAX_SCSI_DEVICES; ii++) {
          ioc->spi_data.dvStatus[ii] =
            MPT_SCSICFG_NEGOTIATE;
@@ -1256,12 +1252,12 @@
          ioc->spi_data.dvStatus[ii] |=
            MPT_SCSICFG_DV_NOT_DONE;
 
-      ddvprintk((MYIOC_s_INFO_FMT
+      dinitprintk((MYIOC_s_INFO_FMT
          "dv %x width %x factor %x saf_te %x\n",
-         ioc->name, driver_setup.dv,
-         driver_setup.max_width,
-         driver_setup.min_sync_factor,
-         driver_setup.saf_te));
+         ioc->name, mpt_dv,
+         mpt_width,
+         mpt_factor,
+         mpt_saf_te));
    }
 
    mpt_scsi_hosts++;
@@ -1477,18 +1473,6 @@
         ": Registered for IOC reset notifications\n"));
    }
 
-#ifdef MODULE
-   dinitprintk((KERN_INFO MYNAM
-      ": Command Line Args: dv=%d max_width=%d "
-      "factor=0x%x saf_te=%d\n",
-      dv, width, factor, saf_te));
-
-   driver_setup.dv = (dv) ? 1 : 0;
-   driver_setup.max_width = (width) ? 1 : 0;
-   driver_setup.min_sync_factor = factor;
-   driver_setup.saf_te = (saf_te) ? 1 : 0;;
-#endif
-
    if(mpt_device_driver_register(&mptscsih_driver,
      MPTSCSIH_DRIVER) != 0 ) {
       dprintk((KERN_INFO MYNAM
@@ -3165,6 +3149,18 @@
    dinitprintk((MYIOC_s_INFO_FMT "initTarget bus=%d id=%d lun=%d hd=%p\n",
          hd->ioc->name, bus_id, target_id, lun, hd));
 
+   /*
+    * If the peripheral qualifier filter is enabled then if the target reports a 0x1
+    * (i.e. The targer is capable of supporting the specified peripheral device type
+    * on this logical unit; however, the physical device is not currently connected
+    * to this logical unit) it will be converted to a 0x3 (i.e. The target is not
+    * capable of supporting a physical device on this logical unit). This is to work
+    * around a bug in th emid-layer in some distributions in which the mid-layer will
+    * continue to try to communicate to the LUN and evntually create a dummy LUN.
+   */
+   if (mpt_pq_filter && dlen && (data[0] & 0xE0))
+      data[0] |= 0x40;
+   
    /* Is LUN supported? If so, upper 2 bits will be 0
    * in first byte of inquiry data.
    */
@@ -5161,7 +5157,7 @@
    }
    ddvprintk((MYIOC_s_NOTE_FMT "DV: Basic test on id=%d completed OK.\n", ioc->name, id));
 
-   if (driver_setup.dv == 0)
+   if (mpt_dv == 0)
       goto target_done;
 
    inq0 = (*pbuf1) & 0x1F;
diff -uarN drivers/message/fusion-3.01.19/mptscsih.h drivers/message/fusion/mptscsih.h
--- drivers/message/fusion-3.01.19/mptscsih.h   2005-03-22 10:43:07.000000000 -0700
+++ drivers/message/fusion/mptscsih.h   2005-03-22 10:50:14.000000000 -0700
@@ -91,12 +91,4 @@
 #define MPTSCSIH_MIN_SYNC               0x08
 #define MPTSCSIH_SAF_TE                 0
 
-struct mptscsih_driver_setup
-{
-        u8      dv;
-        u8      max_width;
-        u8      min_sync_factor;
-        u8      saf_te;
-};
-
 #endif


This new kernel gives me these nice results:
Code:

meister:~# hdparm -tT /dev/sda

/dev/sda:
 Timing buffer-cache reads:   1780 MB in  2.00 seconds = 887.92 MB/sec
 Timing buffered disk reads:  196 MB in  3.03 seconds =  64.67 MB/sec
Back to top
View user's profile Send private message
ekool
n00b
n00b


Joined: 14 Sep 2003
Posts: 8

PostPosted: Sat Mar 26, 2005 10:34 pm    Post subject: Reply with quote

Zeek,

I'm curious as to your hardware setup for these results:

Code:

tux freeipmi # dbench -c /usr/share/dbench/client_oplocks.txt 15
15 clients started
   0     62477  788.75 MB/secc
Throughput 788.74 MB/sec 15 procs


I've run all the patches and get the following:

Code:

root@wretched ekool # hdparm -tT /dev/sda

/dev/sda:
 Timing cached reads:   3096 MB in  2.00 seconds = 1547.46 MB/sec
 Timing buffered disk reads:  198 MB in  3.01 seconds =  65.79 MB/sec


Code:

root@wretched ekool # dbench -c /usr/share/dbench/client_oplocks.txt 15
15 clients started
   0     62477  373.14 MB/sec
Throughput 373.135 MB/sec 15 procs


And the biggest problem I have is that the sql-bench suite shows this system to be slower then single processor systems on SATA drives, as opposed to this 15k RPM SCSI drive (non raid)

Read performance seems to be fine, however write performance seems to be very lacking.

The above tests were run on 2.6.12-rc1 with the patch posted in this thread.
Back to top
View user's profile Send private message
uprooter
n00b
n00b


Joined: 06 Mar 2005
Posts: 70

PostPosted: Sat Dec 12, 2009 7:11 pm    Post subject: 2.6.31 same hardware same problem. Reply with quote

This thread is from 2005, 5 years later..
I have gentoo-sources 2.6.31 installed on IBM eserver 336. I'm not using the raid1 capabilities of the LSI53C1030
hdparm shows
Code:
hdparm -Tt /dev/sda

/dev/sda:
 Timing cached reads:   1754 MB in  2.00 seconds = 877.71 MB/sec
 Timing buffered disk reads:  112 MB in  3.01 seconds =  37.19 MB/sec

also tried on sdb, same results.
It's almost 3 times slower than it should be.
Why ?
Back to top
View user's profile Send private message
uprooter
n00b
n00b


Joined: 06 Mar 2005
Posts: 70

PostPosted: Sat Dec 12, 2009 9:47 pm    Post subject: Reply with quote

Hey I just found that my disks are detected wrong
There are ulta320 (FAST-160) and the detection shows FAST-20
Code:

csi: waiting for bus probes to complete ...
Nov 26 23:35:16 localhost kernel: scsi 2:0:0:0: Direct-Access     IBM      IC35L146UCDY10-0 S27N PQ: 0 ANSI: 3
Nov 26 23:35:16 localhost kernel: target2:0:0: Beginning Domain Validation
Nov 26 23:35:16 localhost kernel: target2:0:0: Domain Validation skipping write tests
Nov 26 23:35:16 localhost kernel: target2:0:0: Ending Domain Validation
Nov 26 23:35:16 localhost kernel: target2:0:0: FAST-20 WIDE SCSI 40.0 MB/s ST (50 ns, offset 8)
Nov 26 23:35:16 localhost kernel: scsi 2:0:1:0: Direct-Access     IBM      IC35L146UCDY10-0 S27N PQ: 0 ANSI: 3
Nov 26 23:35:16 localhost kernel: target2:0:1: Beginning Domain Validation
Nov 26 23:35:16 localhost kernel: target2:0:1: Domain Validation skipping write tests
Nov 26 23:35:16 localhost kernel: target2:0:1: Ending Domain Validation
Nov 26 23:35:16 localhost kernel: target2:0:1: FAST-20 WIDE SCSI 40.0 MB/s ST (50 ns, offset 8)
Nov 26 23:35:16 localhost kernel: scsi 2:0:8:0: Processor         IBM      25P3495a S320  1 1    PQ: 0 ANSI: 2
Nov 26 23:35:16 localhost kernel: target2:0:8: Beginning Domain Validation
Nov 26 23:35:16 localhost kernel: target2:0:8: Ending Domain Validation
Nov 26 23:35:16 localhost kernel: target2:0:8: asynchronous
Nov 26 23:35:16 localhost kernel: BIOS EDD facility v0.16 2004-Jun-25, 2 devices found
Nov 26 23:35:16 localhost kernel: sd 2:0:0:0: [sda] 286749610 512-byte hardware sectors: (146 GB/136 GiB)
Nov 26 23:35:16 localhost kernel: sd 2:0:1:0: [sdb] 286749610 512-byte hardware sectors: (146 GB/136 GiB)
Nov 26 23:35:16 localhost kernel: sd 2:0:1:0: [sdb] Write Protect is off
Nov 26 23:35:16 localhost kernel: sd 2:0:0:0: [sda] Write Protect is off
Nov 26 23:35:16 localhost kernel: sd 2:0:0:0: [sda] Mode Sense: cb 00 00 08
Nov 26 23:35:16 localhost kernel: sd 2:0:1:0: [sdb] Mode Sense: cb 00 00 08
Nov 26 23:35:16 localhost kernel: sd 2:0:0:0: [sda] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA
Nov 26 23:35:16 localhost kernel: sd 2:0:1:0: [sdb] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA
Back to top
View user's profile Send private message
eccerr0r
Watchman
Watchman


Joined: 01 Jul 2004
Posts: 9601
Location: almost Mile High in the USA

PostPosted: Wed Jan 13, 2010 6:48 pm    Post subject: Reply with quote

Did you try an older kernel?
Unfortunately I can't run x86_64 code on my machine with the onboard 53c1030, but with 2.6.29 it seems to work fine (NON RAID):
Code:
mughi ~ # hdparm -tT /dev/sda

/dev/sda:
 Timing cached reads:   2594 MB in  2.00 seconds = 1297.03 MB/sec
 Timing buffered disk reads:  254 MB in  3.02 seconds =  84.08 MB/sec
mughi ~ # uname -a
Linux mughi 2.6.29 #1 SMP Tue Mar 24 19:21:51 MDT 2009 ia64 31 GenuineIntel GNU/Linux
mughi ~ # lspci|grep SCSI
03:01.0 SCSI storage controller: LSI Logic / Symbios Logic 53c1030 PCI-X Fusion-MPT Dual Ultra320 SCSI (rev 08)
03:01.1 SCSI storage controller: LSI Logic / Symbios Logic 53c1030 PCI-X Fusion-MPT Dual Ultra320 SCSI (rev 08)

Code:
scsi target1:0:0: Beginning Domain Validation
scsi target1:0:0: Ending Domain Validation
scsi target1:0:0: FAST-160 WIDE SCSI 320.0 MB/s DT IU QAS RTI WRFLOW PCOMP (6.25 ns, offset 127)
sd 1:0:0:0: [sda] 143638992 512-byte hardware sectors: (73.5 GB/68.4 GiB)
sd 1:0:0:0: [sda] Write Protect is off


Might also want to check your cables to make sure they're good.
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Gentoo on AMD64 All times are GMT
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum