Reminder to self that this is an excellent part of a dashboard for someone with Alzheimer’s disease: [WayBack] Hoe laat het is: hoe laat is het? Zo laat is het.
–jeroen
Archive for 2021
Hoe laat het is: hoe laat is het? Zo laat is het.
Posted by jpluimers on 2021/09/10
Posted in Alzheimer's disease, LifeHacker, Power User | Leave a Comment »
Some links to post about ESXi 6 and ESXi 7 storage and storage speed issues
Posted by jpluimers on 2021/09/09
For my link archive (most via [Wayback] sata very slow after ESXi 6.7 update – Google Search):
- Thick versus thick provisioned VMDKs: [Wayback/Archive.is] Solved: ESXi 6.7 – Windows 10 Pro VM – poor disk performan… – VMware Technology Network VMTN
- PERC firmware and driver versions were too old: [Wayback] Solved: ESXi 6.0U3 disk performance extremely slow – VMware Technology Network VMTN
- Use VMotion to move things around: [Wayback] [SOLVED] ESXi 6.7 slow ssd read rate when moving VM – Spiceworks
- GUI operations might be capped, try
dd: [Wayback] datastore – How can I speed up file tranfers on local storage in vSphere 5? – Server Fault - An older Dell OpenManage Server Administrator version slowing things down: [Wayback] Performance issue with vSphere 6.7U3 on Dell servers – vInfrastructure Blog
- Disable the native driver so the Lynx driver gets selected [Wayback] ESXi 6.5 Storage Performance Issues and Fix – VIRTUALIZATION IS LIFE!
- Passing through the Lynx PCIe to FreeNAS on a SuperMicro X10 based board (and showing how to list PCI devices; I use
lspci | grep -E -i "SAS|SATA|SCSI|RAID|SSD"for this): [Wayback] Freenas on ESXI – AHCI Passthrough – Intel C220 Lynx Point Controller | TrueNAS Community - Another FreeNAS passthrough of a Lynx PCIe device showing
esxcli storage core adapter listto ist controllers together with drivers, thenlspci -n(I tend to uselspci -v | lessfor this) to find which one to set to passthrough: [Wayback] Kaperschip: AHCI controller passthrough with a Supermicro Xeon-D motherboard
- Passing through the Lynx PCIe to FreeNAS on a SuperMicro X10 based board (and showing how to list PCI devices; I use
- The stock
vmw_ahcibeing slow,whereas the Patsburg driver isn’t (this is supposed to be fixed in ESXi 6.7): [Wayback/Archive.is] Incredibly slow disk performance – ESXi 6.5 : vmware- [Wayback] Solved: Very slow speed on SSD – VMware Technology Network VMTN
ESXi 6.5 includes a new native driver (vmw_ahci) for SATA AHCI controllers, but that introduces performance problems with a lot of controllers and/or disks.
Try to disable the native driver and revert to the older sata-ahci driver by running
esxcli system moduleset--enabled=false--module=vmw_ahciin an ESXi shell. Reboot the host to make the change effective.
- [Wayback] Solved: Very slow speed on SSD – VMware Technology Network VMTN
- In specific situations, snapshots can make things very slow [Wayback] windows 10 1809 slow – VMware Technology Network VMTN.
- Not all drivers are installed when upgrade is done through the wrong command: [Wayback] NVMe SSDs are not recognized anymore after upgrading to VMware ESXi 7 – ivobeerens.nl
- Detailed steps to fix this: [Wayback] Important – NVMe SSD not found after upgrading to ESXi 7.0
- These devices do work:
- Samsung SSD 970 EVO 1TB NVMe
- Samsung SSD 950 PRO 512GB NVMe
- Some consumer NVMe devices no longer supported, workaround is to use the old driver (which might not work for future upgrades): [Wayback] Quick Tip – Crucial NVMe SSD not recognized by ESXi 6.7 & 7.0
- Some NVMe devices (even of the same brand/type) suddenly not being visible any more: [Wayback] ESXi 6.7U2 not seeing all the NVMe, only 2 of 12 -… – VMware Technology Network VMTN
Two takeaways already:
- “get a VMUG licence, they are not that expensive and you get to lab up all the latest software.”
- when testing features (in stead of hardware support), nesting is good, though unsupported:
–jeroen
Share this:
Posted in ESXi6, ESXi6.5, ESXi6.7, ESXi7, Power User, Virtualization, VMware, VMware ESXi | Leave a Comment »
PowerShell: recovering from corrupt empty *.nupkg files after a disk was accidentally full during update
Posted by jpluimers on 2021/09/09
When you do a choco upgrade all --yes on a system that – during upgrade – becomes low on disk space, you can end up with a lot of empty .nupkg files.
For those package, Chocolatey will not recognise they are installed any more.
The fix is this:
- increase disk space so at least 5 gigabytes is free
- split the choco upgrade process so it checks before each upgrade that this diskspace is indeed free
- list all choco .nupkg files of length zero ordered from oldest to newest
- for each package, delete the .nupkg file if it exists, then force install it with the
--forceparameter before the--yesparameter like in
choco install --force --yes chocolatey - when all packages have been done, then
choco upgrade --all --yes
I wrote a few PowerShell scripts assisting me in cleaning up the mess.
choco-list-installed.bat
- [WayBack] windows – How to list Chocolatey packages already installed and newer version available from the command line – Super User
:: https://superuser.com/questions/890251/how-to-list-chocolatey-packages-already-installed-and-newer-version-available-fr choco list --localonly %*
choco-show-installed-package-names.bat
:: `--limit-output` does not show Chocolatey version header and count footer. :: `--id-oonly` omits the version number, so you only get the package name choco list --local-only --limit-output --id-only
choco-show-installed-package-names-and-versions.bat
:: `--limit-output` does not show Chocolatey version header and count footer. choco list --local-only --limit-output %*
choco-reinstall-empty-nupkg-by-names.ps1
- [WayBack] Powershell – Finding 0-byte Files | Another computer blog
- [WayBack] windows – Where is the Chocolatey installation path? – Stack Overflow:
There is an environment variable set on installation,
ChocolateyInstall, which is set toC:\Chocolateyby default in versions of Chocolatey less than 0.9.8.27. After that, this defaults toC:\ProgramData\Chocolatey.NOTE: By default, the
C:\ProgramDatafolder on Windows is hidden. You will either need to enable hidden files and folders throughFolder Options | Viewor you can navigate directly to the path shown above by copy/pasting directly into the Windows Explorer address bar.In version 0.9.9 of Chocolatey, it actively moves from the old folder location to the new one.
- [WayBack] string – Powershell concatenate an Environment variable with path – Stack Overflow
A convenient way to obtain the string value rather than the dictionary entry (which is technically what
Get-ChildItemis accessing) is to just use the variable syntax:$Env:USERPROFILErather thanGet-ChildItem Env:USERPROFILE.$localpath = "$env:USERPROFILE\some\path"…
Also, the
Join-Pathcmdlet is a good way to combine two parts of a path.$localpath = Join-Path $env:USERPROFILE 'some\path'
<# https://learningpcs.blogspot.com/2009/12/powershell-finding-0-byte-files.html Zero length .nupkg files sorted by oldest first. These are packages that choco will not show and likekly need a forced reinstall. Choco does remember the version that was installed (so not all the choco config is hosed). - https://stackoverflow.com/questions/28235388/where-is-the-chocolatey-installation-path/28239451#28239451 - https://stackoverflow.com/questions/41047123/powershell-concatenate-an-environment-variable-with-path/41047343#41047343 /#> $LibPath = Join-Path $env:ChocolateyInstall 'lib' $NuPkgExtension = 'nupkg' $NupkgFilter = "*.$NuPkgExtension" ## Remove the empty .nupkg files for each argument $args | ForEach-Object { $PackageName = $_ Write-Output "Deleting any empty $PackageName.$NuPkgExtension under $LibPath :" Get-ChildItem -Path $LibPath -Recurse -Filter $NupkgFilter | Where-Object { ($_.Length -eq 0) -and ($_.BaseName -eq $PackageName) } | Sort-Object LastWriteTime | ForEach-Object { $PackageFullName = $_.FullName Write-Output "Deleting $PackageFullName" Remove-Item $PackageFullName } } ## Force install the chocolatey package for each argument $args | ForEach-Object { $PackageName = $_ Write-Output "Installing $PackageName with Chocolatey:" choco install --force --yes $PackageName }
Link lists
Some more links that helped me solve this:
- [WayBack] Powershell – Finding 0-byte Files | Another computer blog
- [WayBack] CommandsList · chocolatey/choco Wiki · GitHub
- [WayBack] How to get a list of packages from one machine and install in another with Chocolatey? – Stack Overflow
- [WayBack] Create a NuGet package using nuget.exe CLI | Microsoft Docs
- [WayBack] windows – Where is the Chocolatey installation path? – Stack Overflow
- [WayBack] string – Powershell concatenate an Environment variable with path – Stack Overflow
Some links on errors I encountered while recovering from this:
- Checksum errors like[WayBack] (sysinternals) checksum error · Issue #756 · chocolatey-community/chocolatey-coreteampackages · GitHub are often caused by the chocolatey package downloading the most recent installer despite the package version. Two solutions:
- Pass
--ignorechecksumtochoco --install(see [WayBack] CommandsInstall · chocolatey/choco Wiki · GitHub) - First uninstall using the
--forceparameter
[Archive.is] Chocolatey Software | Sysinternals 2019.6.29Sysinternals Suite is going to be installed in ‘C:\ProgramData\chocolatey\lib\sysinternals\tools’
File appears to be downloaded already. Verifying with package checksum to determine if it needs to be redownloaded.
Error – hashes do not match. Actual value was ‘A510C31C2CC591A16F342E7CBA5DC8409EAF08C9B56729CF132C95C69E196787’.
Downloading sysinternals
from ‘https://download.sysinternals.com/files/SysinternalsSuite.zip’
Progress: 100% – Completed download of C:\Users\devCrPhoneDebug\AppData\Local\Temp\2\chocolatey\sysinternals\2018.12.27\SysinternalsSuite.zip (23.51 MB).
Download of SysinternalsSuite.zip (23.51 MB) completed.
Error – hashes do not match. Actual value was ‘A510C31C2CC591A16F342E7CBA5DC8409EAF08C9B56729CF132C95C69E196787’.
ERROR: Checksum for ‘C:\Users\devCrPhoneDebug\AppData\Local\Temp\2\chocolatey\sysinternals\2018.12.27\SysinternalsSuite.zip’ did not meet ‘b14466c6bf3be216ea71610a3f455030e791cd5ad1b42a283886194205d176b0’ for checksum type ‘sha256’. Consider passing the actual checksums through with –checksum –checksum64 once you validate the checksums are appropriate. A less secure option is to pass –ignore-checksums if necessary.
The install of sysinternals was NOT successful.
Error while running ‘C:\ProgramData\chocolatey\lib\sysinternals\tools\chocolateyInstall.ps1’.
See log for details.Chocolatey installed 0/1 packages. 1 packages failed.
See the log for details (C:\ProgramData\chocolatey\logs\chocolatey.log).
- Pass
- Packages that cannot be found at all:[WayBack] “imagemagick not installed. The package was not found with the source(s) listed” – Google Search
- This means that Chocolatey cannot find a dependency, but will not tell you which one. It also happens during package testing:[WayBack] imagemagick v7.0.9.7 – Failed – Package Tests Results · GitHub
- Solve this by fixing all other empty
.nupkgfiles first, which will give you an idea on the potential missing dependencies. Retry by forcing reinstall each dependency.
- Solve this by fixing all other empty
- This means that Chocolatey cannot find a dependency, but will not tell you which one. It also happens during package testing:[WayBack] imagemagick v7.0.9.7 – Failed – Package Tests Results · GitHub
- Dependencies that cannot be found, which can be caused by more empty
.nupkgfiles. Example: [WayBack] Unable to resolve dependency · Issue #206 · chocolatey/choco · GitHub- Solve this by each time a dependency is not found, include on the
choco-reinstall-empty-nupkg-by-names.ps1command, then retry.
- Solve this by each time a dependency is not found, include on the
–jeroen
Share this:
Posted in Chocolatey, COBOL, Development, Power User, PowerShell, PowerShell, Scripting, Software Development, Windows, Windows 10 | Leave a Comment »
Opening shell folders from the command-prompt
Posted by jpluimers on 2021/09/09
I knew I could run shell:startup and similar shortcuts from the Explorer address bar or the Windows-R “run” prompt.
First I learned that via [WayBack] tablet – How to set Google Chrome to automatically open up and in full screen – Super User.
Then via [WayBack] “shell:startup” – Google Search, I found [WayBack] Location of the Startup folder in Windows 10.
It took a while before I realised you can also run them from the command-prompt, batch-files or PowerShell scripts prepending them with start:
start shell:startup
That one will open a new explorer window in the user startup folder from either the command-prompt, a batch file or PowerShell script..
The shell: shortcuts can contain spaces. So for instance there is shell:common startup that opens the common startup folder.
Starting it from the command prompt, batch file or PowerShell script is different: because of the spaces you will get the error on the right unless you add double quotes:
start "shell:common statartup"
All shell: commands that you can run in the same way: double quotes work for both the ones requiring spaces and the simple ones nor requiring spaces.
Virtually each new Windows version (even most Windows 10 major builds) gets new shell: commands.
A good source with an up-to-date and historically accurate of shell: commands list is at [WayBack] Shell Commands to Access the Special Folders in Windows 10/8/7/Vista/XP » Winhelponline,
You can get the current list by recursively enumerating the HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Explorer\FolderDescriptions registry key, which consists of a list of Explorer folder GUIDs having Name, ParentFolder and RelativePath value names.
–jeroen
Share this:
Posted in Batch-Files, Console (command prompt window), Development, Power User, Scripting, Software Development, Windows | Leave a Comment »
Chocolatey parameter order: `–yes` becomes before `–force`
Posted by jpluimers on 2021/09/08
Not sure why, bit if you want to force install a package, answering yes to all prompts, the chocolaty parameter order needs to be --yes --force instead of --force --yes.
This works:
choco install --yes --force git.install --params "/GitAndUnixToolsOnPath /NoGitLfs /SChannel /NoAutoCrlf /WindowsTerminal"
This fails:
choco install --force --yes git.install --params "/GitAndUnixToolsOnPath /NoGitLfs /SChannel /NoAutoCrlf /WindowsTerminal"
–jeroen
Share this:
Posted in Chocolatey, Development, DevOps, Power User, Scripting, Software Development, Windows | Leave a Comment »
PowerShell OS Support Matrix – mohitgoyal.co
Posted by jpluimers on 2021/09/08
By now, probably newer versions have come out, but this should give a rough indication of the 2019 state of [WayBack] PowerShell OS Support Matrix – mohitgoyal.co:
For 5.1 and lower, you can find the prerequisites in [WayBack] Windows PowerShell System Requirements – PowerShell | Microsoft Docs.
–jeroen
Share this:
Posted in .NET, CommandLine, Development, PowerShell, PowerShell, Scripting, Software Development | Leave a Comment »
Most network protocols are TCP based, so be aware ping uses ICMP and traceroute UDP
Posted by jpluimers on 2021/09/08
Interesting thread: [WayBack] SwiftOnSecurity sur Twitter : “I had this issue in my prev company network. QoS will drop ICMP and you’ll chase your tail. If you want to find out if a network service works, test the service. If you want to know if TCP works, use TCP.… “
So:
tcppingandtcptraceroutefor the win!- remember that some protocols, rely on ICMP or UDP, so ensure these work on your network tool (do not QoC them away!)
[WayBack] Zimmie on Twitter: “It is perhaps worth noting: traceroute does not generally use ICMP. Instead, it uses this horrific UDP port range with a different port for every probe at every hop. 30 hops? That’s 90 different UDP ports you just tried. Makes its value questionable at the best of times.…”
Related:
- [WayBack] tcptraceroute(1) – Linux man page
- [WayBack] traceroute(8) – Linux man page
- [WayBack] ping(8) – Linux man page
- [WayBack] tcpping/tcpping at master · deajan/tcpping · GitHub
–jeroen
Share this:
Posted in *nix, *nix-tools, Communications Development, Development, Power User, TCP, UDP | Leave a Comment »
“A parameter cannot be found that matches parameter name ‘PassThru'” – likely your powershell version is too old
Posted by jpluimers on 2021/09/07
If you get [WayBack] “A parameter cannot be found that matches parameter name ‘PassThru'” as PowerShell error, then likely the PowerShell version is too old to support -PassThru, which likely means you have are running pre-Windows 10 version.
PowerShell 3 (introduced in 2012) added the -PassThru parameter that allowed to chain multiple commands from one list pipe.
Another reason for the error might be that the command you use does not support the -PassThru parameter.
To check which commandlets support -PassThru, use the below command (the output is from a Windows 8.1 machine running PowerShell 4.0).
Share this:
Posted in .NET, CommandLine, Development, PowerShell, PowerShell, Software Development | Leave a Comment »
checkdigits onTwitter : “CTRL+S is the new carriage return. Type line of code, CTRL+S to save, start new line of code. #NotParanoid 😬…”
Posted by jpluimers on 2021/09/07
After bumping into yet another occasion where a tool did not properly save after saving: [WayBack] checkdigits on Twitter : “CTRL+S is the new carriage return. Type line of code, CTRL+S to save, start new line of code. #NotParanoid 😬…”
in a response to [WayBack] Dare Obasanjo on Twitter: “No one I know who works on software for a living trusts it to do anything important. That people assume software can be trusted for important things like voting or driving cars when I can’t even trust my follower count is correct is the biggest trick we’ve played as an industry.… “
Share this:
Posted in Development, Software Development | Leave a Comment »
Tricks used by software developers to https://127.0.0.1
Posted by jpluimers on 2021/09/07
Long interesting thread at [WayBack] Thread by @sleevi_: “@SwiftOnSecurity So, some history: It used to be folks would get certs for “localhost”, just like they would from “webmail”, despite no CA e […]”
In 2019, applications were still using tricks (including shipping private keys!) to “securely” access https://127.0.0.1 on some port.
This should have stopped in 2015, but hadn’t. I wonder how bad it still is today.
Related:
- [WayBack] SwiftOnSecurity on Twitter: “Me: Threat-hunting rare DNS lookups in a corporate network. Confluence: … “
- [WayBack] Tavis Ormandy on Twitter: “Yes, it happens sometimes, as soon as someone pulls out the key the CA is required to revoke it. They probably did it to avoid mixed-content warnings, as you can probably guess… it’s not the correct solution. Anyone using this app is vulnerable to trivial MITM 😣… https://t.co/Dqpe5Z9hUu”
- [WayBack] Tavis Ormandy on Twitter: “I pulled out the certificates here, they do seem valid at first glance, not sure why they didn’t get logged in CT. 😆 https://t.co/LyB26oaW5S… https://t.co/DUNIByLSTl”
- [WayBack] Tavis Ormandy on Twitter: “It seems like Atlassian are a CVE CNA, you can ask them to assign a CVE if you like! (seriously, this is a real vulnerability) … “
- [WayBack] SwiftOnSecurity sur Twitter : “Thank you Tavis, I’ve always wondered what it would be like to have a CVE (even if you granted me the courtesy and I didn’t really do much). Thank you a lot. I’ve emailed them and asked for a CVE.… “
- [WayBack] Tavis Ormandy on Twitter: “This turned out to be a real vulnerability! 😮 The certificate was issued by @digicert, who are now required to revoke it. It was issued before mandatory CT, so didn’t show up in …. See … for context.…”
- [WayBack] Ryan Sleevi on Twitter: “So, some history: It used to be folks would get certs for “localhost”, just like they would from “webmail”, despite no CA ever having validated the name. They just relied on pinky promises to be good. Luckily, browsers forbid that … “
- [WayBack] Ryan Sleevi on Twitter: “Look at the dates on those. Yes, it’s a things CAs used to do, and they had to be dragged kicking and screaming into not doing it (and even then, *many* ignored/“oopsied” the requirement to revoke). I regret to inform you it didn’t stop until 2015/2016 – … “
- [WayBack] Guidance on Internal Names – CAB Forum
- [Archive.is] HTTPS encryption on the web – Google Transparency Report: atlassian-domain-for-localhost-connections-only.com
- [Archive.is] HTTPS encryption on the web – Google Transparency Report
- SubjectC=AU, O=Atlassian Pty Ltd, L=Sydney, ST=New South Wales, CN=atlassian-domain-for-localhost-connections-only.com
- Serial NumberA:3E:93:53:0E:74:53:AE:CB:40:BA:20:10:12:F8:FB
- IssuerC=US, O=DigiCert Inc, OU=www.digicert.com, CN=DigiCert SHA2 High Assurance Server CAValidity11 May 2017 — 15 May 2020
- The domain is (at the time of writing, so hopefully that is now “was”) used by the [WayBack] Administering the Atlassian Companion App – Atlassian Documentation
- At the time of writing, the interactive Google DNS showed the domain pointing to localhost [WayBack] Google Public DNS:
Result for atlassian-domain-for-localhost-connections-only.com/A with DNSSEC validation:
{ "Status": 0, "TC": false, "RD": true, "RA": true, "AD": false, "CD": false, "Question": [ { "name": "atlassian-domain-for-localhost-connections-only.com.", "type": 1 } ], "Answer": [ { "name": "atlassian-domain-for-localhost-connections-only.com.", "type": 1, "TTL": 1620, "data": "127.0.0.1" } ] } - [Archive.is/WayBack] SSL Server Test: atlassian-domain-for-localhost-connections-only.com (Powered by Qualys SSL Labs)
Assessment failed: IP address is from private address space (RFC 1918)
Share this:
Posted in Communications Development, Development, DNS, HTTP, Internet, Power User, Software Development, TCP, TLS | Leave a Comment »
The Wiert Corner - irregular stream of stuff 