Jump to content
Hoopster

[-rc15e, 16b, 17b] IPMI Issues

47 posts in this topic Last Reply

Recommended Posts

Updated my backup server to RC15e and lost video output over IPMI.  Video appears at beginning of boot process and I see the logo as BIOS posts, the unRAID menu and most of the boot process also appear. At some point before booting up unRAID finishes, the screen goes black.  Server comes up fine, but, there is no video output via IPMI.  It has worked fine on every version up through RC14. 

 

I rolled back to RC14 and there were no problems.  Video output via IMPI is never lost and comes up to the unRAID login prompt.  Went back to 15e and the problem again occurs.  The server is now running RC14.

 

Diagnostics from 15e attached in case they are helpful. 

 

 

backupnas-diagnostics-20171205-2203.zip

Edited by Hoopster

Share this post


Link to post

Also having problem with lost video over IPMI for my Backup and Test servers, both ASRock boards.

My Main server (Supermicro) is working but the color scheme looks strange.

Main server (Supermicro).png

Share this post


Link to post
1 minute ago, johnnie.black said:

For me IPMI is working fine on Supermicro X9 series, on X11 series letters turn blue.

 

5a281165e11f8_Screenshot2017-12-0614_01_08.png.65c730e5b879b7b9214fa070c9d83d8c.png

My X11DPi-NT is pink.

Share this post


Link to post
13 minutes ago, johnnie.black said:

For me IPMI is working fine on Supermicro X9 series, on X11 series letters turn blue.

 

5a281165e11f8_Screenshot2017-12-0614_01_08.png.65c730e5b879b7b9214fa070c9d83d8c.png

 

It appears the common factor for those of us with no video via IPMI with RC15e is ASRock motherboards.  Supermicro boards show weird colors on some boards.

 

I recently updated the BMC to version 00.28.00 on my ASRock board as they finally fixed the compatibility issue with recent versions of JAVA; however, I doubt that is the problem since even after that update IPMI video was fine on RC14. Linux kernel changes?

Edited by Hoopster

Share this post


Link to post
 

It appears the common factor for those of us with no video via IPMI with RC15e is ASRock motherboards.  Supermicro boards show weird colors on some boards.

 

Whatever the problem is it's most likely related to the new kernel, not unRAID.

 

 

Share this post


Link to post
3 minutes ago, johnnie.black said:

 

Whatever the problem is it's most likely related to the new kernel, not unRAID.

 

 

Agreed.  IPMI video issues would point towards Linux kernel. 

Share this post


Link to post
51 minutes ago, bonienl said:

Anything in your syslog which points to IPMI errors?

 

 

I could see no IPMI-related errors in the syslog.  Here is everything in the syslog that references IPMI directly.  All looks normal relative to IPMI:

 

...
Dec  5 21:59:45 BackupNAS kernel: ipmi message handler version 39.2
...
Dec  5 21:59:45 BackupNAS kernel: ipmi_si IPI0001:00: ipmi_si: probing via ACPI
Dec  5 21:59:45 BackupNAS kernel: ipmi_si IPI0001:00: [io  0x0ca2] regsize 1 spacing 1 irq 0
Dec  5 21:59:45 BackupNAS kernel: ipmi_si: Adding ACPI-specified kcs state machine
Dec  5 21:59:45 BackupNAS kernel: Linux agpgart interface v0.103
Dec  5 21:59:45 BackupNAS kernel: IPMI System Interface driver.
Dec  5 21:59:45 BackupNAS kernel: ipmi_si: probing via SPMI
Dec  5 21:59:45 BackupNAS kernel: ipmi_si: SPMI: io 0xca2 regsize 1 spacing 1 irq 0
Dec  5 21:59:45 BackupNAS kernel: (NULL device *): SPMI-specified kcs state machine: duplicate
...
Dec  5 21:59:45 BackupNAS kernel: ipmi_si: Trying ACPI-specified kcs state machine at i/o address 0xca2, slave address 0x0, irq 0
Dec  5 21:59:45 BackupNAS kernel: ipmi_si IPI0001:00: Found new BMC (man_id: 0x00c1d6, prod_id: 0xaabb, dev_id: 0x20)
Dec  5 21:59:45 BackupNAS kernel: ipmi_si IPI0001:00: IPMI kcs interface initialized
...

This is from the syslog generated last night right after upgrading to RC15e with the latest Linux kernel.

Edited by Hoopster

Share this post


Link to post

Done some testing on my Test server (unRAID3, ASRock - E3C226D2I) and included my log.

The video diapers at row 640: Dec  6 17:39:59 unRAID3 kernel: ipmi_si IPI0001:00: ipmi_si: probing via ACPI

Row 639 is the last that is shown before the video signal is gone.

Log enclosed….

Test server ( ASRock - E3C226D2I).txt

Share this post


Link to post

Firefox Web browser

5 minutes ago, limetech said:

What client are you guys using, supermicro's crappy java-based tool or something else?

 

Firefox browser 57.0.1 (32 bit) and Java 8 Update 152.

Edited by Leifgg

Share this post


Link to post
3 minutes ago, limetech said:

What client are you guys using, supermicro's crappy java-based tool or something else?

I'm using the java based ipmiview, and it still works fine for the X9 series, only the X11 has the weird color change.

Share this post


Link to post
On 12/6/2017 at 10:30 AM, limetech said:

What client are you guys using, supermicro's crappy java-based tool or something else?

 

I don't know how crappy Supermico's java-based IPMI tool is but ASRock's has had its issues.  It just quit working a while ago with any version of Java greater than 8.121.  To their credit, about two-weeks ago, ASRock finally updated the BMC  driver to make it compatible with later versions of Java up to and including the latest 8.151.

 

I am accessing my backup server (ASRock E3C226D2I) via Firefox 57.0 (64-bit) using the ASRock Java IPMI interface with Java 8.151

Edited by Hoopster

Share this post


Link to post
29 minutes ago, Leifgg said:

Done some testing on my Test server (unRAID3, ASRock - E3C226D2I) and included my log.

 

The video diapers at row 640: Dec  6 17:39:59 unRAID3 kernel: ipmi_si IPI0001:00: ipmi_si: probing via ACPI

 

Row 639 is the last that is shown before the video signal is gone.

 

Log enclosed….

 

Test server ( ASRock - E3C226D2I).txt

 

I have the same motherboard in my backup server and see the same results.  It looks like video disappears over IPMI at the point of probing via ACPI.

Share this post


Link to post
3 minutes ago, limetech said:

Last patch I could find in this area is here:

 

That seems to work for me.

Dec  6 09:25:43 vesta kernel: ipmi message handler version 39.2
...
Dec  6 09:25:43 vesta kernel: ipmi_si: probing via SMBIOS
Dec  6 09:25:43 vesta kernel: ipmi_si: SMBIOS: io 0xca2 regsize 1 spacing 1 irq 0
Dec  6 09:25:43 vesta kernel: ipmi_si: Adding SMBIOS-specified kcs state machine
Dec  6 09:25:43 vesta kernel: ipmi_si IPI0001:00: ipmi_si: probing via ACPI
Dec  6 09:25:43 vesta kernel: ipmi_si IPI0001:00: [io  0x0ca2] regsize 1 spacing 1 irq 0
Dec  6 09:25:43 vesta kernel: ipmi_si dmi-ipmi-si.0: Removing SMBIOS-specified kcs state machine in favor of ACPI
Dec  6 09:25:43 vesta kernel: ipmi_si: Adding ACPI-specified kcs state machine
Dec  6 09:25:43 vesta kernel: IPMI System Interface driver.
Dec  6 09:25:43 vesta kernel: ipmi_si: probing via SPMI
Dec  6 09:25:43 vesta kernel: ipmi_si: SPMI: io 0xca2 regsize 1 spacing 1 irq 0
Dec  6 09:25:43 vesta kernel: (NULL device *): SPMI-specified kcs state machine: duplicate
Dec  6 09:25:43 vesta kernel: ipmi_si: Trying ACPI-specified kcs state machine at i/o address 0xca2, slave address 0x20, irq 0
Dec  6 09:25:43 vesta kernel: ahci 0000:00:11.4: version 3.0
Dec  6 09:25:43 vesta kernel: ahci 0000:00:11.4: AHCI 0001.0300 32 slots 4 ports 6 Gbps 0xf impl SATA mode
Dec  6 09:25:43 vesta kernel: ahci 0000:00:11.4: flags: 64bit ncq pm led clo pio slum part ems apst 
Dec  6 09:25:43 vesta kernel: ipmi_si IPI0001:00: The BMC does not support clearing the recv irq bit, compensating, but the BMC needs to be fixed.
...
Dec  6 09:25:43 vesta kernel: ipmi_si IPI0001:00: Found new BMC (man_id: 0x002a7c, prod_id: 0x0857, dev_id: 0x20)
Dec  6 09:25:43 vesta kernel: ipmi_si IPI0001:00: IPMI kcs interface initialized

Though SMBIOS is not present in some of the other reports here.

In my case both java tool and browser work.

Share this post


Link to post

hey guys!

 

i'm on an ASRock Rack EPC612D8A-TB board. it has an VGA connector for IPMI (which also is set for being the primary display) and there is a display connected to it. IPMI has it's own dedicated network port. the board is updated to all latest firmware / BMC and up until (including) RC14 the output on the (real) display was "correctly" in white text on black background.

 

watched yesterday the whole re-boot process with RC15 via IPMI*. i could follow all the way and i didn't get disconnected or anything else – so it works. text/background were (and still are) in black/white (via IPMI).

but on the real display connected to VGA the text color is now blue (background still black). everything else seems to be ok.

 

a grep'ing for 'IPMI' results in these few lines:

Dec  6 18:14:56 unRAID kernel: ipmi message handler version 39.2
Dec  6 18:14:56 unRAID kernel: ipmi_si IPI0001:00: ipmi_si: probing via ACPI
Dec  6 18:14:56 unRAID kernel: ipmi_si IPI0001:00: [io  0x0ca2] regsize 1 spacing 1 irq 0
Dec  6 18:14:56 unRAID kernel: ipmi_si: Adding ACPI-specified kcs state machine
Dec  6 18:14:56 unRAID kernel: IPMI System Interface driver.
Dec  6 18:14:56 unRAID kernel: ipmi_si: probing via SPMI
Dec  6 18:14:56 unRAID kernel: ipmi_si: SPMI: io 0xca2 regsize 1 spacing 1 irq 0
Dec  6 18:14:56 unRAID kernel: ipmi_si: Trying ACPI-specified kcs state machine at i/o address 0xca2, slave address 0x0, irq 0
Dec  6 18:14:56 unRAID kernel: ipmi_si IPI0001:00: Found new BMC (man_id: 0x00c1d6, prod_id: 0xaabb, dev_id: 0x20)
Dec  6 18:14:56 unRAID kernel: ipmi_si IPI0001:00: IPMI kcs interface initialized

* on a separate bloody windows VM, because they suck so much and no solution on Mac is available, which works. and boy i tried a hell lot of stuff. actually i got it to the point, where the remote display would be displayed, but just a second before it craps out with an error '...redirecting floppy controller...' and the java programm gets killed. supermicro has a java utility for linux, which does work on mac (not officially) and uses standard IPMI commands – but here i couldn't get remote display working (maybe with much tinkering of these parameters, not sure and no nerve for that).

Edited by s.Oliver

Share this post


Link to post

No issues here on a X9 motherboard. I have another X9 and X8, but have yet to upgrade them.

Share this post


Link to post

If you are having IPMI video issues, are you booting in 'Normal' mode or 'GUI Mode' (from the boot menu)?

 

Try toggling the mode and see if that changes anything please.

Share this post


Link to post
21 hours ago, limetech said:

If you are having IPMI video issues, are you booting in 'Normal' mode or 'GUI Mode' (from the boot menu)?

 

Try toggling the mode and see if that changes anything please.

 

I have been booting in "normal" mode.  I will reinstall rc15e again and try GUI mode on reboot.  I will edit this post with the results.

 

UPDATE:  After reinstalling rc15e and rebooting, I selected GUI Mode from the unRAID boot menu.  The IMPI screen showed the boot process up to the same point at which video disappears in "normal" mode.  The IPMI console screen again went black for a while; however, the GUI login screen did appear later and I was able to login to the GUI and all appears normal in GUI mode via IPMI.  It is only console output that is lost at some point via IPMI in both normal and GUI modes.

 

I generally prefer Normal mode over GUI mode for the console output and CLI.  Of course, I can get to the CLI via PuTTY and the GUI via a browser, but, IPMI is my console on this server.

 

You are on to something with your question about normal vs. GUI mode.

 

 

Edited by Hoopster

Share this post


Link to post

I have an ASRock E3C226D2 mobo with the Nov 2017 BMC update (it was finally updated to remove a bunch of Java security prompts, yay)

 

I do not have a monitor connected to the box, I access it through the ASRock Java IPMI app. (I spent several hours this morning trying to get SuperMicro's IPMI View Windows app to show the SOL console, but was not successful. At this point the Java app is the only IPMI option.)

 

I had no issues with IPMI in 6.3.5 or rc14. I skipped rc15 but installed rc16 so I could test the microcode updates.  Just mentioning this to say that the November BMC update is not the source of the problems as it worked fine up through rc14.

 

In addition to IPMI, I can connect through the serial port using Putty as described here:
 

Interestingly, the serial console cuts out at the same time as the ASrock IPMI java console, although it comes back with a login prompt a short while later:

pps pps1: new PPS source ptp1
igb 0000:03:00.0: added PHC on eth1
igb 0000:03:00.0: Intel(R) Gigabit Ethernet Network Connection
igb 0000:03:00.0: eth1: (PCIe:2.5Gb/s:Width x1) bc:5f:f4:e7:9a:c3
igb 0000:03:00.0: eth1: PBA No: 001300-000
igb 0000:03:00.0: Using MSI-X interrupts. 4 rx queue(s), 4 tx queue(s)
ipmi_si IPI0001:00: IPMI kcs interface initialized
IPMI SSIF Interface driver

unRAID Server OS version: 6.4.0_rc16b
 IPv4 address: 192.168.10.51
 IPv6 address: not set

Tower login: 

Note that everything drops right after loading the "IPMI SSIF Interface driver".  The full putty recording is attached (love that feature!), maybe it can be matched up with the syslog for additional clues?

 

 

 

---


As an aside, you will also see a bunch of errors that look like this in the syslog:

Dec 16 10:55:24 Tower emhttpd: error: get_device_size, 1421: No medium found (123): open: /dev/sdd
Dec 16 10:55:24 Tower emhttpd: device /dev/sdd problem getting size

That is described in more detail here:
 


 

putty-tower.txt

tower-diagnostics-20171216-1104.zip

Share this post


Link to post

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.