Re: Dell SAS5 Performance Issue
Bjoern A. Zeeb wrote: On Tue, 27 Mar 2007, Ivan Voras wrote: Richard Tector wrote: Is there any information I could obtain that would be of use to you in tracking this one down? You could maybe contact the driver's author(s) directly and ask them. See the AUTHORS section of the man page for details. you should have done your homework before saying this and you might have noticed something;-) To just add a me too; same kind of machine. I have been in contact with people the last three days mostly gathering information for them. And yes, the integrated curses based controller utility you can start during boot does not seem to have options to adjust caching behavior. At least I had not seen and option for that either. Unfortunately this machine will be shipped of the day after tomorrow so I won't be able to test anything new. I'll follow this thread and in case I can help with anything - let me know. /bz -- Bjoern A. Zeebbzeeb at Zabbadoz dot NeT ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED] Hey guys any updates on this? I'm experiencing the same issue win2kx64 environment slow I/O write (read about 65MB/sec) in RAID 1. There is no way to enable caching in the boot BIOS ... anyone hear plans for them to update a version that will allow us to configure these options? -- View this message in context: http://www.nabble.com/Dell-SAS5-Performance-Issue-tf3474648.html#a11065717 Sent from the freebsd-stable mailing list archive at Nabble.com. ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
dmose wrote: Bjoern A. Zeeb wrote: On Tue, 27 Mar 2007, Ivan Voras wrote: Richard Tector wrote: Is there any information I could obtain that would be of use to you in tracking this one down? You could maybe contact the driver's author(s) directly and ask them. See the AUTHORS section of the man page for details. you should have done your homework before saying this and you might have noticed something;-) To just add a me too; same kind of machine. I have been in contact with people the last three days mostly gathering information for them. And yes, the integrated curses based controller utility you can start during boot does not seem to have options to adjust caching behavior. At least I had not seen and option for that either. Unfortunately this machine will be shipped of the day after tomorrow so I won't be able to test anything new. I'll follow this thread and in case I can help with anything - let me know. /bz -- Bjoern A. Zeeb bzeeb at Zabbadoz dot NeT ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED] Hey guys any updates on this? I'm experiencing the same issue win2kx64 environment slow I/O write (read about 65MB/sec) in RAID 1. There is no way to enable caching in the boot BIOS ... anyone hear plans for them to update a version that will allow us to configure these options? Are you asking about FreeBSD or Windows 2000? Scott ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
Scott Long-2 wrote: dmose wrote: Bjoern A. Zeeb wrote: On Tue, 27 Mar 2007, Ivan Voras wrote: Richard Tector wrote: Is there any information I could obtain that would be of use to you in tracking this one down? You could maybe contact the driver's author(s) directly and ask them. See the AUTHORS section of the man page for details. you should have done your homework before saying this and you might have noticed something;-) To just add a me too; same kind of machine. I have been in contact with people the last three days mostly gathering information for them. And yes, the integrated curses based controller utility you can start during boot does not seem to have options to adjust caching behavior. At least I had not seen and option for that either. Unfortunately this machine will be shipped of the day after tomorrow so I won't be able to test anything new. I'll follow this thread and in case I can help with anything - let me know. /bz -- Bjoern A. Zeeb bzeeb at Zabbadoz dot NeT ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED] Hey guys any updates on this? I'm experiencing the same issue win2kx64 environment slow I/O write (read about 65MB/sec) in RAID 1. There is no way to enable caching in the boot BIOS ... anyone hear plans for them to update a version that will allow us to configure these options? Are you asking about FreeBSD or Windows 2000? Scott ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED] Either or (actually Windows2003server x64) I figured the write cache is not an OS specific setting, but rather controlled via the host adapters internal BIOS. So I can't see any way to enable this via the bios.. -- View this message in context: http://www.nabble.com/Dell-SAS5-Performance-Issue-tf3474648.html#a11066546 Sent from the freebsd-stable mailing list archive at Nabble.com. ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
dmose wrote: Scott Long-2 wrote: dmose wrote: Bjoern A. Zeeb wrote: On Tue, 27 Mar 2007, Ivan Voras wrote: Richard Tector wrote: Is there any information I could obtain that would be of use to you in tracking this one down? You could maybe contact the driver's author(s) directly and ask them. See the AUTHORS section of the man page for details. you should have done your homework before saying this and you might have noticed something;-) To just add a me too; same kind of machine. I have been in contact with people the last three days mostly gathering information for them. And yes, the integrated curses based controller utility you can start during boot does not seem to have options to adjust caching behavior. At least I had not seen and option for that either. Unfortunately this machine will be shipped of the day after tomorrow so I won't be able to test anything new. I'll follow this thread and in case I can help with anything - let me know. /bz -- Bjoern A. Zeeb bzeeb at Zabbadoz dot NeT ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED] Hey guys any updates on this? I'm experiencing the same issue win2kx64 environment slow I/O write (read about 65MB/sec) in RAID 1. There is no way to enable caching in the boot BIOS ... anyone hear plans for them to update a version that will allow us to configure these options? Are you asking about FreeBSD or Windows 2000? Scott ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED] Either or (actually Windows2003server x64) I figured the write cache is not an OS specific setting, but rather controlled via the host adapters internal BIOS. So I can't see any way to enable this via the bios.. It's not controllable through the BIOS. There are Windows apps that will allow you to control it, but you should contact Dell and LSI about those, not FreeBSD. FreeBSD now has an option for handling it as well. Scott ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
I've been trying to get one in my lab. I also am completely saturated with two jobs and a new infant (which is why I'm responding to this at 0100) and was trying to get a box in my lab that evidenced te behaviour. If I get stuck when I actually get time to chase this, I'll ask- thanks. On 4/24/07, J. Martin Petersen [EMAIL PROTECTED] wrote: Matthew Jacob wrote: Is there any news on the performance of this card? I personally have not been able to reproduce the problem. It seems to occur whether in Integrated Raid or not. It seems to be related to specific backplanes and drives. It's an important problem to solve I agree. We have a HP Proliant DL140 g3 that exhibits this (or a somewhat related) problem, to which we can give you remote access (including remote KVM) if that helps? Cheers, Martin ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
Matthew Jacob wrote: I've been trying to get one in my lab. I also am completely saturated with two jobs and a new infant (which is why I'm responding to this at 0100) and was trying to get a box in my lab that evidenced te behaviour. If I get stuck when I actually get time to chase this, I'll ask- thanks. SNIP I have one sat on my desk, if you would like I can ship it to you? Tom ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
Matthew Jacob wrote: Is there any news on the performance of this card? I personally have not been able to reproduce the problem. It seems to occur whether in Integrated Raid or not. It seems to be related to specific backplanes and drives. It's an important problem to solve I agree. We have a HP Proliant DL140 g3 that exhibits this (or a somewhat related) problem, to which we can give you remote access (including remote KVM) if that helps? Cheers, Martin ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
On Tue, 17 Apr 2007, Matthew Jacob wrote: Hi, Is there any news on the performance of this card? I personally have not been able to reproduce the problem. It seems to occur whether in Integrated Raid or not. It seems to be related to specific backplanes and drives. It's an important problem to solve I agree. I had done some research a few weeks back and just found the reference again. There is a fujitsu-siemens manual[1], and 2.4.8 Disk write caching gives some hitns on why things are the way they are. I had also found references, that Windows people have effectively seen the same problem. My conclusion is - that's nothing that can be changes with software easily for the Dell PE 860 with that special chip. another 0.002ct References: [1] http://manuals.fujitsu-siemens.com/serverbooks/content/manuals/english/ir-sas-ug-en.pdf -- Bjoern A. Zeeb bzeeb at Zabbadoz dot NeT ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
Richard Tector wrote: Ivan Voras wrote: Richard Tector wrote: I'm suffering from very slow write performance on a Dell PowerEdge 860 with the SAS5/IR controller (mpt driver) running either 6.2-RELEASE or 6.2-STABLE with sources from yesterday. The controller hosts 2 Western Digital 320GB SATA disks in a RAID1 configuration. Reads approach 65MB/s however writes appear extremely slow, in the region of 6-7MB/s with a dd and a blocksize of 1MB all the way down to about 300KB/s while extracting a ports snapshot. It was suggested to me that perhaps write caching has been disabled on the controller however no options exist within the BIOS configuration to view/adjust *any* caching options. You looked in the controller's BIOS, not motherboard's, right? Indeed I did. There should be at least a write through vs write back switch... Correct, there should be options, but there aren't. The controller BIOS has very few options at all in fact. No, but can you post the relevant bits for the controller from dmesg? Sure: mpt0: LSILogic SAS/SATA Adapter port 0xec00-0xecff mem 0xfe9fc000-0xfe9f,0xfe9e-0xfe9e irq 16 at device 8.0 on pci2 mpt0: [GIANT-LOCKED] mpt0: MPI Version=1.5.12.0 mpt0: mpt_cam_event: 0x16 mpt0: Unhandled Event Notify Frame. Event 0x16 (ACK not required). mpt0: mpt_cam_event: 0x12 mpt0: Unhandled Event Notify Frame. Event 0x12 (ACK not required). mpt0: mpt_cam_event: 0x12 mpt0: Unhandled Event Notify Frame. Event 0x12 (ACK not required). mpt0: mpt_cam_event: 0x16 mpt0: Unhandled Event Notify Frame. Event 0x16 (ACK not required). [...snip...] da0 at mpt0 bus 0 target 0 lun 0 da0: Dell VIRTUAL DISK 1028 Fixed Direct Access SCSI-5 device da0: 300.000MB/s transfers, Tagged Queueing Enabled da0: 305175MB (624998400 512 byte sectors: 255H 63S/T 38904C) I have just stumbled across this problem on 4 PE860's and 2 PE840's. I have been through the BIOS of the card and found no information about caching in any of the menus. I then decided to take the card out and could not see any place to attach a cache battery backup unit, I could also not see any ram chips on the card. Is there any news on the performance of this card? Tom ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
Is there any news on the performance of this card? I personally have not been able to reproduce the problem. It seems to occur whether in Integrated Raid or not. It seems to be related to specific backplanes and drives. It's an important problem to solve I agree. ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
Matthew Jacob wrote: On 3/27/07, Richard Tector [EMAIL PROTECTED] wrote: Matthew Jacob wrote: It's been seen before but the fix isn't know yet. On 3/27/07, Richard Tector [EMAIL PROTECTED] wrote: I'm suffering from very slow write performance on a Dell PowerEdge 860 with the SAS5/IR controller (mpt driver) running either 6.2-RELEASE or 6.2-STABLE with sources from yesterday. The controller hosts 2 Western Digital 320GB SATA disks in a RAID1 configuration. Reads approach 65MB/s however writes appear extremely slow, in the region of 6-7MB/s with a dd and a blocksize of 1MB all the way down to about 300KB/s while extracting a ports snapshot. It was suggested to me that perhaps write caching has been disabled on the controller however no options exist within the BIOS configuration to view/adjust *any* caching options. Has anyone else experienced this issue? If so how did they resolve it? Also, is there any way to toggle the caching settings from FreeBSD? A perhaps unrealted issue: I've noticed a large number of messages being produced by the mpt driver after boot occuring approximately every 90 seconds. Apr 1 15:51:43 moses kernel: mpt0: mpt_cam_event: 0x14 Apr 1 15:51:43 moses kernel: mpt0: Unhandled Event Notify Frame. Event 0x14 (ACK not required). Apr 1 15:53:19 moses kernel: mpt0: mpt_cam_event: 0x14 Apr 1 15:53:19 moses kernel: mpt0: Unhandled Event Notify Frame. Event 0x14 (ACK not required). Any thoughts anyone? Is there any way to find out what the events correspond to? Regards, Richard ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
On Sun, 1 Apr 2007, Richard Tector wrote: A perhaps unrealted issue: I've noticed a large number of messages being produced by the mpt driver after boot occuring approximately every 90 seconds. Apr 1 15:51:43 moses kernel: mpt0: mpt_cam_event: 0x14 Apr 1 15:51:43 moses kernel: mpt0: Unhandled Event Notify Frame. Event 0x14 (ACK not required). Apr 1 15:53:19 moses kernel: mpt0: mpt_cam_event: 0x14 Apr 1 15:53:19 moses kernel: mpt0: Unhandled Event Notify Frame. Event 0x14 (ACK not required). Any thoughts anyone? Is there any way to find out what the events correspond to? IMHO 0x14 is resync complete. Dud you have a degraded RAID-1? Can you tell me what Index: mpt_cam.c === RCS file: /shared/mirror/FreeBSD/r/ncvs/src/sys/dev/mpt/mpt_cam.c,v retrieving revision 1.52 diff -u -p -r1.52 mpt_cam.c --- mpt_cam.c 11 Mar 2007 01:54:59 - 1.52 +++ mpt_cam.c 1 Apr 2007 16:54:27 - @@ -2236,6 +2236,10 @@ mpt_cam_event(struct mpt_softc *mpt, req CAMLOCK_2_MPTLOCK(mpt); break; } + case MPI_EVENT_IR_RESYNC_UPDATE: + uint8_t resync = (data0 16) 0xff; + mpt_prt(mpt, IR resync update %d completed., resync); + break; case MPI_EVENT_EVENT_CHANGE: case MPI_EVENT_INTEGRATED_RAID: case MPI_EVENT_SAS_DEVICE_STATUS_CHANGE: gives you; especially how far it counts up? -- Bjoern A. Zeeb bzeeb at Zabbadoz dot NeT ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
On 4/1/07, Richard Tector [EMAIL PROTECTED] wrote: Matthew Jacob wrote: On 3/27/07, Richard Tector [EMAIL PROTECTED] wrote: Matthew Jacob wrote: It's been seen before but the fix isn't know yet. On 3/27/07, Richard Tector [EMAIL PROTECTED] wrote: I'm suffering from very slow write performance on a Dell PowerEdge 860 with the SAS5/IR controller (mpt driver) running either 6.2-RELEASE or 6.2-STABLE with sources from yesterday. The controller hosts 2 Western Digital 320GB SATA disks in a RAID1 configuration. Reads approach 65MB/s however writes appear extremely slow, in the region of 6-7MB/s with a dd and a blocksize of 1MB all the way down to about 300KB/s while extracting a ports snapshot. It was suggested to me that perhaps write caching has been disabled on the controller however no options exist within the BIOS configuration to view/adjust *any* caching options. Has anyone else experienced this issue? If so how did they resolve it? Also, is there any way to toggle the caching settings from FreeBSD? A perhaps unrealted issue: I've noticed a large number of messages being produced by the mpt driver after boot occuring approximately every 90 seconds. Apr 1 15:51:43 moses kernel: mpt0: mpt_cam_event: 0x14 Apr 1 15:51:43 moses kernel: mpt0: Unhandled Event Notify Frame. Event 0x14 (ACK not required). Apr 1 15:53:19 moses kernel: mpt0: mpt_cam_event: 0x14 Apr 1 15:53:19 moses kernel: mpt0: Unhandled Event Notify Frame. Event 0x14 (ACK not required). Any thoughts anyone? Is there any way to find out what the events correspond to? MPI spec, instantiated in MPILIB MPI_EVENT_IR_RESYNC_UPDATE internal raid resync update? ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
Bjoern A. Zeeb wrote: On Sun, 1 Apr 2007, Richard Tector wrote: A perhaps unrealted issue: I've noticed a large number of messages being produced by the mpt driver after boot occuring approximately every 90 seconds. Apr 1 15:51:43 moses kernel: mpt0: mpt_cam_event: 0x14 Apr 1 15:51:43 moses kernel: mpt0: Unhandled Event Notify Frame. Event 0x14 (ACK not required). Apr 1 15:53:19 moses kernel: mpt0: mpt_cam_event: 0x14 Apr 1 15:53:19 moses kernel: mpt0: Unhandled Event Notify Frame. Event 0x14 (ACK not required). Any thoughts anyone? Is there any way to find out what the events correspond to? IMHO 0x14 is resync complete. Dud you have a degraded RAID-1? Can you tell me what Index: mpt_cam.c === RCS file: /shared/mirror/FreeBSD/r/ncvs/src/sys/dev/mpt/mpt_cam.c,v retrieving revision 1.52 diff -u -p -r1.52 mpt_cam.c --- mpt_cam.c 11 Mar 2007 01:54:59 - 1.52 +++ mpt_cam.c 1 Apr 2007 16:54:27 - @@ -2236,6 +2236,10 @@ mpt_cam_event(struct mpt_softc *mpt, req CAMLOCK_2_MPTLOCK(mpt); break; } + case MPI_EVENT_IR_RESYNC_UPDATE: + uint8_t resync = (data0 16) 0xff; + mpt_prt(mpt, IR resync update %d completed., resync); + break; case MPI_EVENT_EVENT_CHANGE: case MPI_EVENT_INTEGRATED_RAID: case MPI_EVENT_SAS_DEVICE_STATUS_CHANGE: gives you; especially how far it counts up? Now you come to mention it, I believe it was indeed the array rebuilding. I'll try your patch and force a resync of the mirror tomorrow morning and report back with the output. Would be a useful feature to have. On a side note the array is now marked as Optimal in the BIOS so I assume the following messages, sent about 2 hours later, were various ways of saying Complete? Apr 1 18:31:43 moses kernel: mpt0: mpt_cam_event: 0x14 Apr 1 18:31:43 moses kernel: mpt0: Unhandled Event Notify Frame. Event 0x14 (ACK not required). Apr 1 18:31:43 moses kernel: mpt0: mpt_cam_event: 0x21 Apr 1 18:31:43 moses kernel: mpt0: Unhandled Event Notify Frame. Event 0x21 (ACK not required). Apr 1 18:31:43 moses kernel: mpt0: mpt_cam_event: 0x21 Apr 1 18:31:43 moses kernel: mpt0: Unhandled Event Notify Frame. Event 0x21 (ACK not required). Apr 1 18:31:43 moses kernel: mpt0: mpt_cam_event: 0x15 Apr 1 18:31:43 moses kernel: mpt0: Unhandled Event Notify Frame. Event 0x15 (ACK not required). Apr 1 18:31:43 moses kernel: mpt0: mpt_cam_event: 0x21 Apr 1 18:31:43 moses kernel: mpt0: Unhandled Event Notify Frame. Event 0x21 (ACK not required). Apr 1 18:31:43 moses kernel: mpt0: mpt_cam_event: 0x21 Apr 1 18:31:43 moses kernel: mpt0: Unhandled Event Notify Frame. Event 0x21 (ACK not required). Thank you for your help, Richard ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
On Tue, 27 Mar 2007, Matthew Jacob wrote: On 3/27/07, Richard Tector [EMAIL PROTECTED] wrote: Thank you for the quick reply. Is there any information I could obtain that would be of use to you in tracking this one down? not at the momeny- it's really that I won't have time until april As said, I won't have the chance to test anything from this evening on and I had a few minutes so I added the usual printfs to get the detailed information for the reported mpt_cam_event()s. : grep ^mpt dell_pe860-20070328-02.log mpt0: LSILogic SAS/SATA Adapter port 0xec00-0xecff mem 0xfe9fc000-0xfe9f,0xfe9e-0xfe9e irq 16 at device 8.0 on pci2 mpt0: Reserved 0x100 bytes for rid 0x10 type 4 at 0xec00 mpt0: Reserved 0x4000 bytes for rid 0x14 type 3 at 0xfe9fc000 mpt0: [GIANT-LOCKED] mpt0: MPI Version=1.5.12.0 mpt0: Event on. mpt0: SAS discovery started. mpt0: SAS PHY(0x00) link status: Rate: 3.0 Gbps. mpt0: SAS PHY(0x01) link status: Rate: 3.0 Gbps. mpt0: SAS discovery done. That's also what the controller reports (but I think the driver does not try to re-neg/discover anything itself here?). Most likely that's it from me on the write performance topic. -- Bjoern A. Zeeb bzeeb at Zabbadoz dot NeT ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
Richard Tector wrote: I'm suffering from very slow write performance on a Dell PowerEdge 860 with the SAS5/IR controller (mpt driver) running either 6.2-RELEASE or 6.2-STABLE with sources from yesterday. The controller hosts 2 Western Digital 320GB SATA disks in a RAID1 configuration. Reads approach 65MB/s however writes appear extremely slow, in the region of 6-7MB/s with a dd and a blocksize of 1MB all the way down to about 300KB/s while extracting a ports snapshot. It was suggested to me that perhaps write caching has been disabled on the controller however no options exist within the BIOS configuration to view/adjust *any* caching options. You looked in the controller's BIOS, not motherboard's, right? There should be at least a write through vs write back switch... If you don't have a battery for the SAS controller, get one. Has anyone else experienced this issue? No, but can you post the relevant bits for the controller from dmesg? signature.asc Description: OpenPGP digital signature
Re: Dell SAS5 Performance Issue
It's been seen before but the fix isn't know yet. On 3/27/07, Richard Tector [EMAIL PROTECTED] wrote: I'm suffering from very slow write performance on a Dell PowerEdge 860 with the SAS5/IR controller (mpt driver) running either 6.2-RELEASE or 6.2-STABLE with sources from yesterday. The controller hosts 2 Western Digital 320GB SATA disks in a RAID1 configuration. Reads approach 65MB/s however writes appear extremely slow, in the region of 6-7MB/s with a dd and a blocksize of 1MB all the way down to about 300KB/s while extracting a ports snapshot. It was suggested to me that perhaps write caching has been disabled on the controller however no options exist within the BIOS configuration to view/adjust *any* caching options. Has anyone else experienced this issue? If so how did they resolve it? Also, is there any way to toggle the caching settings from FreeBSD? Regards, Richard Tector ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED] ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
Ivan Voras wrote: Richard Tector wrote: I'm suffering from very slow write performance on a Dell PowerEdge 860 with the SAS5/IR controller (mpt driver) running either 6.2-RELEASE or 6.2-STABLE with sources from yesterday. The controller hosts 2 Western Digital 320GB SATA disks in a RAID1 configuration. Reads approach 65MB/s however writes appear extremely slow, in the region of 6-7MB/s with a dd and a blocksize of 1MB all the way down to about 300KB/s while extracting a ports snapshot. It was suggested to me that perhaps write caching has been disabled on the controller however no options exist within the BIOS configuration to view/adjust *any* caching options. You looked in the controller's BIOS, not motherboard's, right? Indeed I did. There should be at least a write through vs write back switch... Correct, there should be options, but there aren't. The controller BIOS has very few options at all in fact. No, but can you post the relevant bits for the controller from dmesg? Sure: mpt0: LSILogic SAS/SATA Adapter port 0xec00-0xecff mem 0xfe9fc000-0xfe9f,0xfe9e-0xfe9e irq 16 at device 8.0 on pci2 mpt0: [GIANT-LOCKED] mpt0: MPI Version=1.5.12.0 mpt0: mpt_cam_event: 0x16 mpt0: Unhandled Event Notify Frame. Event 0x16 (ACK not required). mpt0: mpt_cam_event: 0x12 mpt0: Unhandled Event Notify Frame. Event 0x12 (ACK not required). mpt0: mpt_cam_event: 0x12 mpt0: Unhandled Event Notify Frame. Event 0x12 (ACK not required). mpt0: mpt_cam_event: 0x16 mpt0: Unhandled Event Notify Frame. Event 0x16 (ACK not required). [...snip...] da0 at mpt0 bus 0 target 0 lun 0 da0: Dell VIRTUAL DISK 1028 Fixed Direct Access SCSI-5 device da0: 300.000MB/s transfers, Tagged Queueing Enabled da0: 305175MB (624998400 512 byte sectors: 255H 63S/T 38904C) Regards, Richard ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
Matthew Jacob wrote: It's been seen before but the fix isn't know yet. On 3/27/07, Richard Tector [EMAIL PROTECTED] wrote: I'm suffering from very slow write performance on a Dell PowerEdge 860 with the SAS5/IR controller (mpt driver) running either 6.2-RELEASE or 6.2-STABLE with sources from yesterday. The controller hosts 2 Western Digital 320GB SATA disks in a RAID1 configuration. Reads approach 65MB/s however writes appear extremely slow, in the region of 6-7MB/s with a dd and a blocksize of 1MB all the way down to about 300KB/s while extracting a ports snapshot. It was suggested to me that perhaps write caching has been disabled on the controller however no options exist within the BIOS configuration to view/adjust *any* caching options. Has anyone else experienced this issue? If so how did they resolve it? Also, is there any way to toggle the caching settings from FreeBSD? Regards, Richard Tector ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED] Thank you for the quick reply. Is there any information I could obtain that would be of use to you in tracking this one down? Regards, Richard ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
Richard Tector wrote: Is there any information I could obtain that would be of use to you in tracking this one down? You could maybe contact the driver's author(s) directly and ask them. See the AUTHORS section of the man page for details. signature.asc Description: OpenPGP digital signature
Re: Dell SAS5 Performance Issue
Bjoern A. Zeeb wrote: you should have done your homework before saying this and you might have noticed something;-) Ah, yes, I see now :) My apologies - I only glanced and compared the e-mail addresses of the posters and the authors and arrived to the obvious conclusion :) signature.asc Description: OpenPGP digital signature
Re: Dell SAS5 Performance Issue
On Tue, 27 Mar 2007, Ivan Voras wrote: Richard Tector wrote: Is there any information I could obtain that would be of use to you in tracking this one down? You could maybe contact the driver's author(s) directly and ask them. See the AUTHORS section of the man page for details. you should have done your homework before saying this and you might have noticed something;-) To just add a me too; same kind of machine. I have been in contact with people the last three days mostly gathering information for them. And yes, the integrated curses based controller utility you can start during boot does not seem to have options to adjust caching behavior. At least I had not seen and option for that either. Unfortunately this machine will be shipped of the day after tomorrow so I won't be able to test anything new. I'll follow this thread and in case I can help with anything - let me know. /bz -- Bjoern A. Zeeb bzeeb at Zabbadoz dot NeT ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Dell SAS5 Performance Issue
not at the momeny- it's really that I won't have time until april On 3/27/07, Richard Tector [EMAIL PROTECTED] wrote: Matthew Jacob wrote: It's been seen before but the fix isn't know yet. On 3/27/07, Richard Tector [EMAIL PROTECTED] wrote: I'm suffering from very slow write performance on a Dell PowerEdge 860 with the SAS5/IR controller (mpt driver) running either 6.2-RELEASE or 6.2-STABLE with sources from yesterday. The controller hosts 2 Western Digital 320GB SATA disks in a RAID1 configuration. Reads approach 65MB/s however writes appear extremely slow, in the region of 6-7MB/s with a dd and a blocksize of 1MB all the way down to about 300KB/s while extracting a ports snapshot. It was suggested to me that perhaps write caching has been disabled on the controller however no options exist within the BIOS configuration to view/adjust *any* caching options. Has anyone else experienced this issue? If so how did they resolve it? Also, is there any way to toggle the caching settings from FreeBSD? Regards, Richard Tector ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED] Thank you for the quick reply. Is there any information I could obtain that would be of use to you in tracking this one down? Regards, Richard ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to [EMAIL PROTECTED]
