From 2f4a99bbe8cd51c854bbb49e1c864124b8b2283e Mon Sep 17 00:00:00 2001 From: jmott85 <51213510+jmott85@users.noreply.github.com> Date: Wed, 31 Mar 2021 06:30:33 -0500 Subject: [PATCH 1/5] Update H310.md Added instructions for disabling third party PCI fan response in iDRAC after changing to non-Dell firmware. --- docs/H310.md | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-) diff --git a/docs/H310.md b/docs/H310.md index ebc9774..4b1c790 100644 --- a/docs/H310.md +++ b/docs/H310.md @@ -73,10 +73,26 @@ flashboot /root/Bootloaders/x64sas2.rom ``` You can now ditch the live images and boot back into your normal system. +## Optional: (Dell) Modify RACADM + +>Warning: Use this at your own risk. Modifying thermal settings can create unforseen circumstances. If you are running your server in a warm environment, it is probably best to leave this alone. + +The LSI firmware is not supported by Dell. This causes the iDRAC to no longer keep track of the drive temperatures. In response, the iDRAC will throw error PCI3018 in the Lifecycle Log and set the fans to a static speed of about 30%. + +To correct this, you can disable the "ThirdPartyPCIFanResponse" feature by using RACADM. The easiest way to do this is to connect to the iDRAC using an SSH client. Once connected, run this command: +``` +racadm set system.thermalsettings.ThirdPartyPCIFanResponse 0 +``` +If you need to revert, run this command instead: +``` +racadm set system.thermalsettings.ThirdPartyPCIFanResponse 1 +``` +>Note: This information was collected from the Dell iDRAC9 User's Guide which can be referenced here: https://dl.dell.com/topicspdf/idrac9-lifecycle-controller-v33-series_users-guide6_en-us.pdf . The relevant information can be found near the bottom of page 57. + ## Optional: Reverting If for some reason you need to revert back to the stock Dell PERC firmware, that's easy. Boot into the FreeDOS live image, and run the following command: ``` 310REVRT ``` That's it! When it finishes, just reboot back to your normal system with the `reboot` command. ->Note: This uses the unmodified latest Dell firmware `20.13.3-0001,A11` extracted from the update EXE found [here](https://www.dell.com/support/home/us/en/04/drivers/driversdetails?driverid=yp0nf&oscode=ws8r2&productcode=poweredge-r720). \ No newline at end of file +>Note: This uses the unmodified latest Dell firmware `20.13.3-0001,A11` extracted from the update EXE found [here](https://www.dell.com/support/home/us/en/04/drivers/driversdetails?driverid=yp0nf&oscode=ws8r2&productcode=poweredge-r720). From 21d8a81c3006993437c05923da18d8115616ce07 Mon Sep 17 00:00:00 2001 From: jmott85 <51213510+jmott85@users.noreply.github.com> Date: Wed, 31 Mar 2021 06:38:20 -0500 Subject: [PATCH 2/5] Update H310.md --- docs/H310.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/H310.md b/docs/H310.md index 4b1c790..459d896 100644 --- a/docs/H310.md +++ b/docs/H310.md @@ -75,9 +75,9 @@ You can now ditch the live images and boot back into your normal system. ## Optional: (Dell) Modify RACADM ->Warning: Use this at your own risk. Modifying thermal settings can create unforseen circumstances. If you are running your server in a warm environment, it is probably best to leave this alone. +>Warning: Use this at your own risk. Modifying thermal settings can cause unforeseen circumstances. If you are running your server in a warm environment, it is probably best to leave this alone. -The LSI firmware is not supported by Dell. This causes the iDRAC to no longer keep track of the drive temperatures. In response, the iDRAC will throw error PCI3018 in the Lifecycle Log and set the fans to a static speed of about 30%. +The LSI firmware is not supported by Dell. This means the iDRAC will no longer keep track of the drive temperatures. This will cause the iDRAC to throw error PCI3018 in the Lifecycle Log and set the fans to a static speed of about 30%. It uses this static fan speed as a failsafe to prevent any disks from possibly overheating. To correct this, you can disable the "ThirdPartyPCIFanResponse" feature by using RACADM. The easiest way to do this is to connect to the iDRAC using an SSH client. Once connected, run this command: ``` From 99f087a307944e1e0328664ea13a5085c28b3253 Mon Sep 17 00:00:00 2001 From: jmott85 <51213510+jmott85@users.noreply.github.com> Date: Mon, 5 Apr 2021 22:15:11 -0500 Subject: [PATCH 3/5] Update perc.md Moved instructions for disabling ThirdPartyPCIFanResponse from H310.md to perc.md Updated wording in several areas to make it easier to read. Added instructions for ipmitool and made that the main solution with RACADM being secondary Changed reference to whitepaper instead of iDRAC user's guide Removed "Optional" and added "Extra" to be more in-line with current standard in document --- docs/perc.md | 50 +++++++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 49 insertions(+), 1 deletion(-) diff --git a/docs/perc.md b/docs/perc.md index fd6bde3..9ce6134 100644 --- a/docs/perc.md +++ b/docs/perc.md @@ -62,6 +62,54 @@ If it displays an **H810 Adapter D1** revision, proceed to the [H810 Full Size ( If it displays anything that doesn't exactly match the above choices, [contact me](mailto:jon@fohdeesha.com?subject=PERC-Unknown) with a screenshot. If you're impatient and pick the "closest one" instead, you'll brick your card. +## Extra: Disable ThirdPartyPCIFanResponse + +>Warning: Use this at your own risk. Modifying thermal settings can cause unforeseen circumstances. If you are running your server in a warm environment, it is probably best to leave this alone. + +The LSI firmware is not supported by Dell. This will cause the iDRAC to no longer keep track of the drive temperatures. This is confirmed with the error PCI3018 in the Lifecycle Log and the fans being set to a static speed of about 30%. The fan speed acts as a failsafe to prevent any disks from possibly overheating. + +To correct this behavior, you can disable the "ThirdPartyPCIFanResponse" feature by using IPMItool or RACADM. IPMItool is built into the live image so this will usually be the easiest option. If you are no longer booted into the live image, use the RACADM option instead to disable via SSH. + +### Option 1: Disable ThirdPartyPCIFanResponse using IPMItool via Linux Shell + +IPMItool is built into the live image. Otherwise, you will need to install IPMItool in a Linux environment first. + +To disable ThirdPartyPCIFanResponse, run the following command in the Linux shell: +``` +ipmitool –I open raw 0x30 0xce 0 0x16 5 0 0 0 5 0 1 0 0 +``` + +To verify if the option is enabled or not, run this command in the Linux shell: +``` +ipmitool –I open raw 0x30 0xce 1 0x16 5 0 0 0 +``` + +If you need to enable the feature again, run this command in the Linux shell: +``` +ipmitool –I open raw 0x30 0xce 0 0x16 5 0 0 0 5 0 0 0 0 +``` + +### Option 2: Disable ThirdPartyPCIFanResponse using RACADM via SSH + +Connect to port 22 on your iDRAC's IP address using your preferred SSH client. + +To disable ThirdPartyPCIFanResponse, run the following command via the SSH client: +``` +racadm set system.thermalsettings.ThirdPartyPCIFanResponse 0 +``` + +To verify if the option is enabled or not, run this command via the SSH client: +``` +racadm getsystem.thermalsettings.ThirdPartyPCIFanResponse +``` + +If you need to enable the feature again, run this command via the SSH client: +``` +racadm set system.thermalsettings.ThirdPartyPCIFanResponse 1 +``` + +>Note: This information was collected from the Dell White Paper "Disabling a Third-Party PCIe Card Cooling Response with Dell PowerEdge Servers" which can be referenced here: https://fohdeesha.com/data/other/perc/ThirdPartyPCIFanResponse.pdf . The relevant information can be found on pages 6 and 7. + ## Extra: Revision Info & Part Numbers The main difference between the B0 and D1 revisions is the D1 will link at PCIe 3.0 speeds, instead of PCIe 2.0. This will almost certainly never cause a bottleneck unless you have every SAS port on the card connected to very fast SSDs that all get hammered at the same time. Even then, you're likely to reach the card's processor limitations before the bus bandwidth limit. The difference with the *P* cards (H710P) is more cache (1GB vs 512MB), but this is totally irrelevant when running the card flashed to IT mode as the cache is not used. @@ -120,4 +168,4 @@ Regardless, since they're the same price used, you may as well grab the D1 revis ### Contributing: The markdown source for these guides is hosted on [**my Github repo.**](https://github.com/Fohdeesha/lab-docu) If you have any suggested changes or additions feel free to submit a pull request. -```Documentation version:``` [ v2.4 (02-23-2021)](https://github.com/Fohdeesha/lab-docu/commits/master) \ No newline at end of file +```Documentation version:``` [ v2.4 (02-23-2021)](https://github.com/Fohdeesha/lab-docu/commits/master) From b765b67956f1e6ba7cc74261bb74e36359f1acaa Mon Sep 17 00:00:00 2001 From: jmott85 <51213510+jmott85@users.noreply.github.com> Date: Mon, 5 Apr 2021 22:17:38 -0500 Subject: [PATCH 4/5] Update perc.md Removed reference to pages since the whole document is relevant. --- docs/perc.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/perc.md b/docs/perc.md index 9ce6134..3937e57 100644 --- a/docs/perc.md +++ b/docs/perc.md @@ -108,7 +108,7 @@ If you need to enable the feature again, run this command via the SSH client: racadm set system.thermalsettings.ThirdPartyPCIFanResponse 1 ``` ->Note: This information was collected from the Dell White Paper "Disabling a Third-Party PCIe Card Cooling Response with Dell PowerEdge Servers" which can be referenced here: https://fohdeesha.com/data/other/perc/ThirdPartyPCIFanResponse.pdf . The relevant information can be found on pages 6 and 7. +>Note: This information was collected from the Dell White Paper "Disabling a Third-Party PCIe Card Cooling Response with Dell PowerEdge Servers" which can be referenced here: https://fohdeesha.com/data/other/perc/ThirdPartyPCIFanResponse.pdf . ## Extra: Revision Info & Part Numbers The main difference between the B0 and D1 revisions is the D1 will link at PCIe 3.0 speeds, instead of PCIe 2.0. This will almost certainly never cause a bottleneck unless you have every SAS port on the card connected to very fast SSDs that all get hammered at the same time. Even then, you're likely to reach the card's processor limitations before the bus bandwidth limit. The difference with the *P* cards (H710P) is more cache (1GB vs 512MB), but this is totally irrelevant when running the card flashed to IT mode as the cache is not used. From b086af6e75d4edc02c4f062867bfeb6529d5dcd0 Mon Sep 17 00:00:00 2001 From: jmott85 <51213510+jmott85@users.noreply.github.com> Date: Mon, 5 Apr 2021 22:18:30 -0500 Subject: [PATCH 5/5] Update H310.md Removed instructions for updating ThirdPartyPCIFanResponse. This has been moved to perc.md --- docs/H310.md | 16 ---------------- 1 file changed, 16 deletions(-) diff --git a/docs/H310.md b/docs/H310.md index 459d896..e3ef975 100644 --- a/docs/H310.md +++ b/docs/H310.md @@ -73,22 +73,6 @@ flashboot /root/Bootloaders/x64sas2.rom ``` You can now ditch the live images and boot back into your normal system. -## Optional: (Dell) Modify RACADM - ->Warning: Use this at your own risk. Modifying thermal settings can cause unforeseen circumstances. If you are running your server in a warm environment, it is probably best to leave this alone. - -The LSI firmware is not supported by Dell. This means the iDRAC will no longer keep track of the drive temperatures. This will cause the iDRAC to throw error PCI3018 in the Lifecycle Log and set the fans to a static speed of about 30%. It uses this static fan speed as a failsafe to prevent any disks from possibly overheating. - -To correct this, you can disable the "ThirdPartyPCIFanResponse" feature by using RACADM. The easiest way to do this is to connect to the iDRAC using an SSH client. Once connected, run this command: -``` -racadm set system.thermalsettings.ThirdPartyPCIFanResponse 0 -``` -If you need to revert, run this command instead: -``` -racadm set system.thermalsettings.ThirdPartyPCIFanResponse 1 -``` ->Note: This information was collected from the Dell iDRAC9 User's Guide which can be referenced here: https://dl.dell.com/topicspdf/idrac9-lifecycle-controller-v33-series_users-guide6_en-us.pdf . The relevant information can be found near the bottom of page 57. - ## Optional: Reverting If for some reason you need to revert back to the stock Dell PERC firmware, that's easy. Boot into the FreeDOS live image, and run the following command: ```