Posted on Leave a comment

Downgrade docker by installing an older version on linux

Sometimes the latest version of software may include bugs that have not been fixed.This is the case for docker on Ubuntu 22.04. The buildkit version prints on stderr instead of stdout and this causes some issues on teamcity. As a solution I wanted to downgrade docker on a previous version. First you will need to delete the existing docker installation. You can do this with the below commands:

sudo apt-get purge -y docker-engine docker docker.io docker-ce docker-ce-cli docker-compose-plugin
sudo apt-get autoremove -y --purge docker-engine docker docker.io docker-ce docker-compose-plugin

sudo rm -rf /var/lib/docker /etc/docker
sudo rm /etc/apparmor.d/docker
sudo groupdel docker
sudo rm -rf /var/run/docker.sock

Based on your OS version you can go under docker downloads and find a specific version. Some older docker version exist for older OS versions but you can install them on later versions also.

Index of linux/ubuntu/dists/ (docker.com)

There could be a case when you have jammy (22.04) but you want to install binaries that were available on bionic version (18.04). In order to install an older version of docker you will need to download the old binaries by navigating inside the specific version then selecting pool and finally stable. Inside the last folder you can find all the architectures that are supported and you should select the appropriate one. For x64 you can select amd64 and then you can find a specific version of the binary.

In order to install specific version you will need to download all the packages that you need (some of them are dependencies of docker-ce) with wget and then install them.

Download:

wget https://download.docker.com/linux/ubuntu/dists/bionic/pool/stable/amd64/docker-scan-plugin_0.17.0~ubuntu-bionic_amd64.deb

Install:

dpkg -i docker-scan-plugin_0.17.0~ubuntu-bionic_amd64.deb
Posted on Leave a comment

npm ERR! ERESOLVE could not resolve dependency

I was working on a project which had a node install task as a part of a docker file build. The build was not successful due to the error shown below. This is a dependency error and can be solved by choosing the right versions on your package-lock.json.

package-lock.json is automatically generated for any operations where npm modifies either the node_modules tree, or package.json. It describes the exact tree that was generated, such that subsequent installs are able to generate identical trees, regardless of intermediate dependency updates.

However there is also another way you can get rid of such cases. If you do have specific dependencies inside your package-lock.json you can proceed with the following steps.

Remove package-lock.json from your project
Remove node_modules folder 
Run npm install 
Posted on Leave a comment

Monitoring Windows service on Azure with Event Viewer and Log Analytics

A Log Analytics workspace is a unique environment for log data from Azure Monitor and other Azure services, such as Microsoft Sentinel and Microsoft Defender for Cloud. Expect from that it can be used also for monitoring combined with Azure alerts given that you create the appropriate query.

The below query can be used to monitor a windows server service by querying log analytics. In more detail it searches for eventID=7036 which indicates the service stopped status.

Query code in Kusto language will return the service name, state and time of the event.

Event
| where TimeGenerated >ago(1h)
| where EventLog  == "System" and EventID ==7036 and Source == "Service Control Manager" 
| parse kind=relaxed EventData with * '<Data Name="param1">' Windows_Service_Name '</Data><Data Name="param2">' Windows_Service_State '</Data>'*
| sort by TimeGenerated desc
| where Windows_Service_Name startswith "Docker Desktop" and Windows_Service_State contains "stopped"
| project Computer, Windows_Service_Name, Windows_Service_State, TimeGenerated

You can use the above query to create a azure alert when a service is found as stopped. As I want to monitor the Docker Desktop service, I will need to use that in the where clause of the query where Windows_Service_Name. The alert logic should indicate when a result is returned as a row in a given timeframe then an alert should be generated. This happens because a row is returned only when the event is captured on the event viewer. This means that the service stopped during the TimeGenerated interval of the query. The frequency of evaluation will be the time on which we want to repeat the log analytics query. For example if we need to search every 5 minutes for a stopped service then we should add 5 minutes there.

Finally the alert will be triggered and inform you about windows stopped services.

Posted on Leave a comment

Curl slack webhook with powershell

The below powershell can be used to trigger a webhook URL for slack. Inside the powershell you can dynamically get variables from powershell using the json notation that is used.

$json = @"
{
    "text": "I am inside $($Env:ComputerName)"
}
"@


if (-not((Get-Service -Name "Appinfo").Status -eq "Running") -or -not((Get-Service -Name "Dhcp").Status -eq "Running")) 
{ 
curl -X POST -H 'Content-type: application/json' --data $json https://hooks.slack.com/services/XXXX/XXXX/XXXX 
}