gmane.linux.hardware.dell.poweredge

DTK 4.2 and Watchdog Timer

Lloyd Brown — 2013-05-24T17:13:23

We've got an in-house developed mechanism for using the hardware
watchdog timer to reboot the node if not responsive within 15 minutes
(900 seconds).  We mostly use ipmitool to set this, and an /etc/cron.d/
entry to reset it back to 900 seconds periodically.

I've been experimenting with DTK 4.2 the last couple of days on some PE
M620s, and for reasons that are not clear, when I do a
"srvadmin-services.sh start", something is periodically resetting my
watchdog parameters to be 480 seconds, and to have no action (rather
than the hard reset we wanted).

Any time I re-set things back to our chosen settings, if the DTK
srvadmin-services are started, it gets set back to 480sec/no-action
within 10-15 seconds.  When I turn the services off
(srvadmin-services.sh stop), it keeps my 900sec/hard-reset settings.

Any idea what's going on here?  I'm getting a little confused, to be
honest.  I haven't figured out which specific service is causing the
problem yet, but it certainly seems to be related to the DTK installatio

updating bios firmware m710HD

Sid Young — 2013-05-24T02:40:57

G'Day all,

I am trying to update the BIOS firmware in a bunch of M710HD with no
success, has anyone done this and how???

Thanks

Sid Young
_______________________________________________
Linux-PowerEdge mailing list
Linux-PowerEdge< at >dell.com
https://lists.us.dell.com/mailman/listinfo/linux-poweredge

Re: RAID 0 vs RAID x

L. A. Walsh, Tlinx Solutions — 2013-05-24T00:10:43






Brent Kimberley wrote:
um...define 'well'?

You can have spectacular failure with any setup given the right
circumstances... of course
that depends on what you mean by spectacular, as well... ;-)

If I had the bays and the disk, i'd probably go raid 10..

As it is, using a raid6, a 2-raid5's and a large raid 50...

Re: Dell PE1950 - PCI Riser Option

Tim Nelson — 2013-05-23T14:27:15

----- Original Message -----

Very interesting. Also odd that the 2950s don't have the same requirement. I have one with both PCI-X, and PCIe risers, working perfectly.

I'll see if I can source the other riser to test.

Thanks!

--Tim

Re: Bad RAID performance on r815+H700

Dennis Jacobfeuerborn — 2013-05-23T12:42:27

Yes, I tried various tweaks with little effect. The only improvement was 
when I activated cut-through I/O by setting the cache to writethrough 
and disabling read-ahead. This raised the iops number in the 6 disk case 
from 12k to 16k. Nice improvement but this still doesn't really get me 
anywhere near the expected numbers.

What perplexes me is that with the number of "spindles" increasing in a 
raid-10 I would expect an *increase* in performance rather than a 
decrease since the iops can be distributed.


Yes, we also have some SuperMicro systems with 9271 controllers here and 
they perform way better.

Regards,
   Dennis

Re: Bad RAID performance on r815+H700

Mark Nelson — 2013-05-23T11:53:52

I've got some R515s with H700 cards I've been having some struggles with 
too, though in a different scenario.

Does it help at all if you use the libaio engine and set iodepth to 
something like 16?  Beyond that, try tweaking the controller settings, 
and various block settings like nr_requests, the scheduler (probably 
want noop for this), and anything else that seems relevant.

In the end, I've not been able to match the kind of performance I can 
get with drives directly connected to an LSI SAS2208 based controller.

Good luck!

Mark

Bad RAID performance on r815+H700

Dennis Jacobfeuerborn — 2013-05-23T11:04:38

Hi,
I'm struggling with configuring a raid-10 with 6 Intel 520 SSDs and I'm 
not sure what the cause of the Problem could be. These are the values I 
get when performing a random write benchmark:

single disk: iops ~ 26k
raid-1: iops ~ 20k
4 disk raid-10: iops ~ 20k
6 disk raid-10: iops ~ 12k

I'm a bit perplexed why adding disks to the raid-10 would make things 
slower rather than faster. I'm using fio to make the tests and this is 
the configuration I use:

[randwrite]
ioengine=sync
rw=randwrite
direct=1
filename=/root/testfile
numjobs=8
group_reporting=1
bs=4k
runtime=30

Any idea what could be going on here? I would expect to get at a minimum 
the same 26k iops with raid-10 that I get with a single disk.

Regards,
    Dennis

Re: Dell PE1950 - PCI Riser Option

Stefan Botter — 2013-05-23T10:14:15

Hi Tim,

On Wednesday, May 22, 2013 09:56:39 PM Tim Nelson wrote:

You need to replace both riser cards. The 1950 is either PCIe or PCI-X, 
no mixed configuration possible.

Greetings,

Stefan

Dell PE1950 - PCI Riser Option

Tim Nelson — 2013-05-22T19:56:39

Yes, not entirely Linux specific, but this system will run CentOS 6.4 x64...

I have a Dell 1950 server with PCIe risers. Due to requirements of using some older/aging hardware, I need to have a PCI slot on the server. Recently, I purchased a replacement riser (the long one) which is PCI-X. If I power up the server with the PCI-X riser, and my card installed, the system hangs indefinitely during POST.

Some quick searching says maybe the PCI-X risers were only compatibile with the Gen III servers, but that doesn't sound right given the Gen I/Gen II systems came with PCI-X at some point.

SO:

-Am I doing something wrong using a PCI-X riser in a non-Gen III 1950?
-Is there a BIOS setting or firmware update required to use this new riser?
-Is it simply possible I have a bad riser?

Thoughts/comments/rants/etc?

Thanks!

--Tim

Re: [..] USB (MOUSE) PROBLEM CENTOS PE2650

Eero Volotinen — 2013-05-22T13:18:16

Hi,

Looks like driver problem, try following:

- update to latest versions of (bios, kernel ..)
- try another mouse (different type of mouse)

--
Eero


2013/5/22 <lefevre62< at >free.fr>

_______________________________________________
Linux-PowerEdge mailing list
Linux-PowerEdge< at >dell.com
https://lists.us.dell.com/mailman/listinfo/linux-poweredge

Re: [..] USB (MOUSE) PROBLEM CENTOS PE2650

2013-05-22T12:44:49

New discovery !
I've plug my mouse/screen and keyboard on back panel, and the mouse is ok under
centos, but...

When i plug a usb mouse on usb port from back panel, no mouse !


&#8220;Dmesg | grep usb&#8221; says :

Usbcore: registered new driver usbfs
Usbcore: registered new driver hub
Usbcore: registered new driver hiddev
Usbcore: registered new driver usbhid
Drivers/usb/input/hid-core.c: v2.6:USBHID&#8230;.
Usb 1-2:new full speed USB device using ohci_hcd and address 1
Usb 1-2: device descriptor read/64, error -110


Idem for address 2, 3, 4,5



---------------------------------------------------------

What dmesg says after plugging usb mouse/usb device? Are USB-ports enabled
on bios?


2013/5/22 <lefevre62 at free.fr>

--

Re: [..] USB (MOUSE) PROBLEM CENTOS PE2650

Eero Volotinen — 2013-05-22T09:39:25

What dmesg says after plugging usb mouse/usb device? Are USB-ports enabled
on bios?


2013/5/22 <lefevre62< at >free.fr>

_______________________________________________
Linux-PowerEdge mailing list
Linux-PowerEdge< at >dell.com
https://lists.us.dell.com/mailman/listinfo/linux-poweredge

[..] USB (MOUSE) PROBLEM CENTOS PE2650

2013-05-22T09:09:38

Hi,


I've a problem on centos on Poweredge 2650.
The mouse (or usb) is not reconized.

How can i investigate ?


Thanks a lot.

Re: idrac7 M620 question

Jaap Dijkshoorn — 2013-05-21T12:55:28

Dear Shine,

Thanks!! that is the command, infact i found it out myself as well, but 
to be complete i did this:

root# /opt/dell/srvadmin/bin/idracadm -r r38n3.con -u root -p 
<yourpassword> set iDRAC.IPMILan.AlertEnable 1
Security Alert: Certificate is invalid - self signed certificate
Continuing execution. Use -S option for racadm to stop execution on 
certificate-related errors.
[Key=iDRAC.Embedded.1#IPMILan.1]
Object value modified successfully

root# /opt/dell/srvadmin/bin/idracadm -r r38n4.con -u root -p 
<yourpassword> eventfilters set -c idrac.alert.system.warning -a none -n 
email
Security Alert: Certificate is invalid - self signed certificate
Continuing execution. Use -S option for racadm to stop execution on 
certificate-related errors.
Set success

root# /opt/dell/srvadmin/bin/idracadm -r r38n4.con -u root -p 
<yourpassword> eventfilters set -c idrac.alert.system.critical -a none 
-n email
Security Alert: Certificate is invalid - self signed certificate
Continuing execution. Use -S option for r

Re: idrac7 M620 question

2013-05-21T12:07:31

Try this command
racadm set idrac.IPMILan.alertEnable enabled

Thanks & Regards

Shine K A
Product Group
Dell | Bangalore Development Center
Phone: +91 80 280 77756
Mob: +91 9448175422


-----Original Message-----
From: linux-poweredge-bounces-Lists On Behalf Of Jaap Dijkshoorn
Sent: Tuesday, May 21, 2013 4:59 PM
To: linux-poweredge-Lists
Subject: Re: [Linux-PowerEdge] idrac7 M620 question

Hi Cupertino,

Thanks, but i already digged through the manual and you can enable a lot but the Main Alert enable config option for racadm cli.

Maybe it is hidden somewhere.. strange thing is you can enable al kind of alerts, but Alerts must be enabled separately looks like it.

On 21-05-13 13:11, cupertino wrote:


--
Met vriendelijke groeten / Kind regards,

Jaap P. Dijkshoorn

| Cluster Computing | SURFsara | Science Park 140 | 1098 XG Amsterdam
| T +31205923000    | jaap.dijkshoorn< at >surfsara.nl | skype: jaap< at >sara.nl
| www.surfsara.nl   | home.sara.nl/~jaapd

Re: idrac7 M620 question

Jaap Dijkshoorn — 2013-05-21T11:29:12

Hi Cupertino,

Thanks, but i already digged through the manual and you can enable a lot 
but the Main Alert enable config option for racadm cli.

Maybe it is hidden somewhere.. strange thing is you can enable al kind 
of alerts, but Alerts must be enabled separately looks like it.

On 21-05-13 13:11, cupertino wrote:

Re: idrac7 M620 question

cupertino — 2013-05-21T11:11:00

quite comprehensive racadm guide can be found here
ftp://ftp.dell.com/Manuals/all-products/esuprt_electronics/esuprt_software/esuprt_remote_ent_sys_mgmt/integrated-dell-remote-access-cntrllr-7-v1.20.20_Reference%20Guide_en-us.pdf 
a quick search for 'alert' revealed many options, though I haven't tried
any yet. 
maybe that helps.


On Tue, 2013-05-21 at 13:05 +0200, Jaap Dijkshoorn wrote:

Re: vdisk and percs 6i <-> h700

cupertino — 2013-05-21T11:07:28

no problem from perc6 -> h700. the other way around it most likely won't
work.

On Tue, 2013-05-21 at 11:37 +0100, lejeczek wrote:

idrac7 M620 question

Jaap Dijkshoorn — 2013-05-21T11:05:28

Hi,

When i am in the DRAC GUI interface (M620) and i go to Alerts you can 
enable or disable Alerts. Which racadm command is the equivalent to 
enable or disable Alerts.

thanks!

vdisk and percs 6i <-> h700

lejeczek — 2013-05-21T10:37:44

dear all

I wonder if anybody's had any experience with moving vdisk 
between perc 6i and h700
what I hopefor I guess is a confirmation that this works, 
array move/migrate between these controllers problem-free
anybody?
regards
_______________________________________________
Linux-PowerEdge mailing list
Linux-PowerEdge< at >dell.com
https://lists.us.dell.com/mailman/listinfo/linux-poweredge

Re: OpenManage 7.2.1 yum repository posted

Thew, Alan — 2013-05-17T11:07:50

Hi,
what can I do on a 2950 with Red Hat 6.4 where I'm trying to upgrade the DRAC firmware (to 1.65)
and I get:

....
Collecting inventory...
FATAL: Module ipmi_msghandler not found.
FATAL: Module ipmi_si not found.

I've used the fix for OMSA but what would people suggest for this problem?

Many thanks