Category Archives: Tech

ESXi 5.5 crashing – Windows Guest won’t load – A simple reminder.

Over the past few days some new guests on my whitebox vmhost (which has been running for almost a year now) started behaving badly! Symptoms listed below, and I’m sure there are were more.  There wasn’t really any pattern to the crashing, it was very random. Things went awry after I added my ninth guest and the host peaked into the 14-16GB of RAM range. My existing guests running PRTG, Zoneminder, VPN and NGINX all were fine… all were stable. Just a few new Windows guests I have been using for testing were crashing. I couldn’t even reinstall windows without a BSOD while loading.

  • VMware ESX unrecoverable error: (vcpu-0)
  • MONITOR PANIC: Unable to decompress PPN from swap slot for VM
  • loading windows starting and crashing Msrpc.sys
  • Windows MMC’s not loading, crashing

I ran memtest86+ on the full 32 gigs with no failures (with only one pass, probably not a good idea) and figured it couldn’t be RAM.

Googling took me down some rabbit holes unloading custom NIC drivers I’ve added. Almost ready to reinstall ESXi. But the issue still persisted.  Finally decided to run memtest  actually in the VM that was having troubles. Errors within seconds. I pulled half the RAM, ran it again… no issues. Looks like it’s bad RAM. RMA time.

So an issue I’ve seen many times back in my Geek Squad days in college- got me. A reminder, even in 2016, don’t rule out bad memory so quick, it would have saved me a few hours tonight.

And finally, the purple screen!

foto_no_exif (3)

bad memory, right?

Certification Authority Event ID 80

After upgrading from a 2003 to 2008 R2 certification authorities I noticed warnings for event ID 80 in the CA logs.  I think I have the fix worked out. Essentially you need to convert global groups to universal then to domain local. Add the CA computer objects, then set some permissions.

Replace the paths with your domain specific information.

2015-11-04_1045

1) On the parent domain,  on a global catalog domain controller (Run from an elevated cmd prompt)

dsmod group "CN=Cert Publishers,CN=Users,DC=domain,DC=company,DC=com" -scope u
dsmod group "CN=Cert Publishers,CN=Users,DC=domain,DC=company,DC=com" -scope l
2) For each child domains, on a global catalog domain controller (Run from an elevated cmd prompt)
dsmod group "CN=Cert Publishers,CN=Users,DC=child,DC=domain,DC=company,DC=com" -scope u
dsmod group "CN=Cert Publishers,CN=Users,DC=child,DC=domain,DC=company,DC=com" -scope l
dsacls "DC=child,DC=domain,DC=company,DC=com" /I:S /G "domain\Cert Publishers":RP;userCertificate
dsacls "DC=child,DC=domain,DC=company,DC=com" /I:S /G "domain\Cert Publishers":WP;userCertificate
dsacls "cn=adminsdholder,cn=system,DC=child,DC=domain,DC=company,DC=com" /G "domain\Cert Publishers":RP;userCertificate
dsacls "cn=adminsdholder,cn=system,DC=child,DC=domain,DC=company,DC=com" /G "domain\Cert Publishers":WP;userCertificate

3) Add the computer objects for your certification authorities to the group  “Cert Publishers” on each domain.
4) Finally, on your certification authorities run the following ((Run from an elevated cmd prompt))

certutil -setreg SetupStatus -SETUP_DCOM_SECURITY_UPDATED_FLAG 
net stop certsvc 
net start certsvc

Whitebox ESXi Host

I’ve been considering  building an ESXi host for home to move my VMs off of Vmware workstation on my Windows 7 Desktop. I wanted something powerful but in a small form factor and uses a relatively low amount of electricity. Well… I think I’m almost there.

I’ll update this as I can with performance but so far I’m happy with the results. Right now I have 4 VMs running including PRTG, OpenVPN, and Zoneminder. So far I am pleased.

I have ESXi 5.5 running off of a flash drive.

Physical Components:
(from Amazon)
Shuttle – SH87R6 (You’ll have to inject the NIC driver for the Realtek hardware into the 5.5 image to use the onboard NIC. So far it seems to work great.)
Intel Core i7-4790S Processor (8M Cache, 3.2 GHz)
4x Seagate 500GB Laptop SATA 7200RPM 32MB Cache 2.5-Inch HDD (configured in RAID 10)
2x PNY XLR8 DDR3 16GB (2x8GB) 1600MHz (PC3-12800)(Four DIMMS total for 32GB)
LSI MegaRAID 256MB SAS RAID Controller MR SAS 8888ELP w/ LSIIBBU05(from eBay, pulled from some HP)

Screen-Shot-2015-04-24-at-5.46.52-PM

 

Screen Shot 2015-04-24 at 5.47.19 PMHardware monitoring looks good
Screen Shot 2015-04-23 at 8.13.23 PM

When backing up my VMs I had an interesting network graph that sort of looks like Michigan

 

Monitoring Ubiquiti APs with PRTG

I have a couple of Ubiquity APs in my home and have an instance of PRTG running. I was looking for a good way to monitor the access points.

Screen Shot 2015-04-04 at 9.03.58 AM

Out of the box Ubquiti supports SNMP and SSH. I wanted a way to monitor how many clients are connected. If you want to monitor the access points configure SNMP on the device, feed the string in to PRTG and the SSH credentials.

Then I wrote a simple shell script below that extracts information from the mca-dump table for connected clients. I’m sure there are different and probably more efficient ways to do this, but this works for me. Essentially the code greps the number and formats the return correctly for PRTG to read, including a status code.

1) Create the directory /var/prtg/scripts/
2) Using VI on the access point create the following file with the code below in /var/prtg/scripts/
3) Save it as wifiusers.sh
4) Add a custom SSH sensor for your access point.(I called mine Wifi Clients)
5) Set the primary channel to “Value (#)”


cmd1=`mca-dump | grep num_sta | tail -n1 | sed 's/[^0-9]*//g`
echo 0:$cmd1:users