Recently, an ubuntu update to the 5.15.0-106 kernel produced the following error from the nvidia driver when it compiled the kernel module:
ERROR: modpost: GPL-incompatible module nvidia.ko uses GPL-only symbol 'rcu_read_unlock_strict'
We typically freeze our driver updates so that we can do them at planned downtimes – the fix was to unhold the held nvidia-driver packages and upgrade to nvidia-driver-550
I needed to replace a drive reported for predictive failure on a Dell 730xd so I ran:
racadm raid replacephysicaldisk:Disk.Bay.3:Enclosure.Internal.0-1:RAID.Integrated.1-1 -dstpd Disk.Bay.10:Enclosure.Internal.0-1:RAID.Integrated.1-1
This will replace a disk in your virtual disk, but won’t update the iDRAC display until the iDRAC is reset, or the machine rebooted. You can verify the job is finished with the lifecycle log, or by using the percli tool to view the topology before doing the reset.
perccli /c0 show
You can get the tool here:
https://www.dell.com/support/article/ca/en/cabsdt1/sln283135/how-to-use-the-poweredge-raid-controller-perc-command-line-interface-cli-utility-to-manage-your-raid-controller?lang=en
You can reset the iDRAC from racadm:
racadm racreset soft
I got the following error after I finished a cut. The problem was that the z-axis had gone too far done, and when it powered up it was already in contact with the z-axis switch.
To fix it, I had to remove the side panel on the side that contains the control panel and adjust the wheel that controlled the z-axis screw, moving the platform up manually. Doing that allowed the cutter to resume normal operation.
I had been asked to create a one-button solution to dial a videoconference on a polycom group series. The unit was on a mobile cart and did not have a fixed IP.
I used the Polycom Intergrator Manual to get the API command needed:
dial addressbook name
Where name is the unique first name of an entry in the address book. I needed to create a serial connection from the button to the polycom serial cable (part number 2457-63542-001) . I first tried an arduino with a MAX3232C transceiver and DB9 connector, but it could not send the signal required. I switched to a raspberry pi instead and a FTDI usb to serial adapter which worked correctly.
Here are the steps you need to follow to do an update to a ESXi5.5 server from the command line:
Taken from the following guide:
https://pubs.vmware.com/vsphere-51/index.jsp?topic=%2Fcom.vmware.vsphere.upgrade.doc%2FGUID-E51C5DB6-F28E-42E8-ACA4-0EBDD11DF55D.html
Note: You must run the following to enable https access for esxcli:
esxcli network firewall ruleset set -e true -r httpClient
1. Find an update profile to use:
esxcli software sources profile list \ --depot=https://hostupdate.vmware.com/software/VUM/PRODUCTION/main/vmw-depot-index.xml | grep 201410
The | is optional but you can grep for later versions.
2. Check with dryrun
esxcli software profile update --dry-run -d \ https://hostupdate.vmware.com/software/VUM/PRODUCTION/main/vmw-depot-index.xml \ -p ESXi-5.5.0-201410001001s-standard Enter Maintenance mode:
esxcli <conn_options> system maintenanceMode set --enable true to enter maintenance mode. |
After all virtual machines on the host have been suspended or migrated, the host enters maintenance mode. You cannot deploy or power on a virtual machine on hosts in maintenance mode.
esxcli <conn_options> system maintenanceMode set --enable false to have a host exit maintenance mode. |
3. Use that profile to update:
esxcli software profile update -d / https://hostupdate.vmware.com/software/VUM/PRODUCTION/main/vmw-depot-index.xml \ -p ESXi-5.5.0-201410001001s-standard
4. Review Event log – you should see an installation message like the following:
Successfully installed 17 VIB(s), removed 16 VIB(s). Please use 'esxcli software profile get' or see log for more detail about the transaction.
The storcli command line can be installed on linux, vmware, windows and macos to provide management of raid devices on booted ibm 3650. The MegaRaid GUI provides a guided wizard for RAID50 disk creation, but the storcli command line options aren’t as clear. It turns out that you must provided an arrangement for the disks in order to create a raid 50 group. For example:
./storcli /c0 add vd r50 name=446SSDRaid50 drives=9:3-8 strip=128 Controller = 0 Status = Failure Description = operation not possible for current RAID level, Invalid RAID level
The above fails because it is missing the pdperarray options, which tells storcli how to arrange the disks:
/opt/StorCLI # ./storcli /c0 add vd type=raid50 name=446SSDRaid50 drives=9:6-11 \ pdperarray=3 strip=128 Controller = 0 Status = Success Description = Add VD Succeeded
Click here
for some further examples of the storcli command.
Following the directions on-line for android and eclipse defaults to an installation which can be run by users in a limited permission environment. This is good, but also means that they the computer they use ends up with a copy of eclipse and the sdk for each user in their home directory space. As an alternative, I tried the following steps to allow students to use the same copy of the sdk and eclipse for development. The instructions below are for windows 7, but an installation on linux or mac os would probably be able to use some of the same steps.
To install the software:
To make available to other users:
I found that when installing Lion from a formatted drive, the ppd files for HP printers aren’t present. You can get them from the previous OS install disc (snow leopard), in the following location (it’s hidden, so use CMD-K in the finder to browse to it):
/Volumes/Mac OS X Install DVD/System/Installation/Packages/
In there you should find a set of all the printer drivers for various printers that could have been installed previously. I recommend using grep to determine which package has the driver you are looking for.
I found a neat trick on-line to get windows 7 to recognize the virtual network card it doesn’t have a default driver for in vmware server 1.0.10. You add the following line to the end of the .vmx config file, and it sees it as an intel e1000 card and installs the correct driver.
ethernet0.virtualDev = "e1000"
This is posted about in a few places, but here is a link to the one I found:
http://social.technet.microsoft.com/Forums/en-US/w7itprohardware/thread/5c74d385-98f9-47c8-962f-b0cfc47abadf
Alan discovered that we can deal with a stale nfs file handle on zeus by using the following command:
umount -l directory
This performs a lazy unmount, which cleans up references to a filesystem after detaching it.