covid 19 – Travelling to Dubai from Berlin with a connection in Amsterdam – do I need a PCR test?

I read the emirates travel guide and it says that passengers arriving from Germany do not need to take the PCR before travel.

I am travelling from Germany but I have connection in Amsterdam with a 1 hour 15 minute layover.

Does this mean I have to take the test?

I could save 60 Euros if I don’t have to.

sql server – Lost of heartbeat and connection with secondary replica

Virtual Enviroment description:

  • WSFC composed by two nodes with Windows server 2016 Standard

  • SQL Server AlwaysOn with synchronous replica and automatic failover

  • SQL Server 2014 (SP3-CU2) 12.0.6214.1

  • Following the cluster thresholds:
    CrossSubnetDelay : 4000
    CrossSubnetThreshold : 40
    PlumbAllCrossSubnetRoutes : 0
    SameSubnetDelay : 2000
    SameSubnetThreshold : 20

  • AG properties:
    LeaseTimeout: 20000
    FailureConditionLevel: 3
    HealthCheckTimeout: 30000
    VerboseLogging: 0

Issue:

RHS.exe process on primary node lost heartbeat with availability group and it initiated the failover. Immediately after the lost of heartbeat, primary node loses the connection with the secondary node and automatic failover fail. Shortly after, hearbeat with primary node works again and primary node takes over the resources again.

Log Details

In chronological order:

  1. Primary node randomly shows signs of disconnection with the secondary replica already in the days leading up to failover:
    ...AlwaysOn Availability Groups connection with secondary database terminated for primary database 'nameofDB'..
  2. The primary node storage randomly shows signs of distress already in the days leading up to failover and just before failover:
    ...Long Sync IO: ... IOs in nonpreemptive mode longer than 1000ms
    FlushCache cleaned up 146070 bufs with 20273 writes in 80387ms...
    SQL Server has encountered 4 occourence(s) of I/O request taking longer than 15 seconds to complete...
  3. RHS.exe lost heartbeat with AG and WSFC send failover request to AG:
    (hadrag) Failure detected, diagnostics heartbeat is lost
    The local replica of availability group ... is preparing to transition to the resolving role..
  4. Shortly after the failover request, the primary replica loses connection with the secondary:
    ...AlwaysOn Availability Groups connection with secondary database terminated for primary database 'nameofDB'..
  5. The listener stops working and there is a checkpoint failure on a specific database:
    One or more recovery units belonging to database .. failed to generate a checkpoint
  6. Eleven seconds after the attempted failover, heartbeat works again and primary replica
    takes over the resources

Questions

Million dollar questions:

  1. why heartbeat was lost?
  2. why connection with secondary replica was lost ?
  3. Any workaround ? increase lease timeout, performance of storage or RAM and CPU can help?

I understand that it is difficult to answer but it can help to only have a strategy or hypotheses that can help me find the problem

linux – CUPS – Network printer – Connection error: Permission denied, The printer is not responding

I’m trying to troubleshoot a printing issue, but first I need to describe my environment (it’s a bit non-standard).

I have a HP 1102 USB printer, connected to a Ubuntu 20.04 system which runs CUPS and shares the printer to other systems in the LAN. Printing from other systems (Ubuntu 18.04, Ubuntu 20.04, Windows 10) works fine via IPP.

I have a client which runs Android 10, and on top of it it runs Linux Deploy with Linux 18.04 in a chroot in the same LAN. The linux system has XFCE and CUPS, and I’ve configured the printer the same way as for other systems.

When I try to print (with local cups set to debug), this is what I get for my job:

D (25/Nov/2020:10:12:01 +0200) (Job 8) Connecting to 192.168.1.13:631
D (25/Nov/2020:10:12:01 +0200) (Job 8) Connecting to printer.
D (25/Nov/2020:10:12:01 +0200) (Job 8) Connection error: Permission denied
E (25/Nov/2020:10:12:01 +0200) (Job 8) The printer is not responding.
D (25/Nov/2020:10:12:31 +0200) (Job 8) Connecting to 192.168.1.13:631
D (25/Nov/2020:10:12:31 +0200) (Job 8) Connecting to printer.
D (25/Nov/2020:10:12:31 +0200) (Job 8) Connection error: Permission denied
E (25/Nov/2020:10:12:31 +0200) (Job 8) The printer is not responding.

I’m doing a packet capture on the cups server side and there are no packets when it says it’s trying to connect to the printer. Connecting manually (telnet 192.168.1.13 631 opens a socket and TCP communication works).

Something must be wrong inside the CUPS running on the Android linux chroot, but I can’t figure out what. The local print job looks fine, file permissions for the /var/spool/cups directory look fine, cups is running as root.

I’m not sure where to look further for troubleshooting…

I’m watching android logcat trying to see if there is a problem at the same time, and it’s this:

11-25 10:28:29.179  2980  3091 W Netd    : No subsystem found in netlink event
11-25 10:28:31.012 20789 20789 I printers.cgi: type=1400 audit(0.0:785): avc: denied { ioctl } for path="socket:(151451)" dev="sockfs" ino=151451 ioctlcmd=0x8933 scontext=u:r:magisk:s0 tcontext=u:r:magisk:s0 tclass=unix_dgram_socket permissive=1

So – I’m thinking cups is trying to create a socket and android is denying it? Any ideas if it’s a file-based socket, what it’s name would be and where it would be located in the filesystem?

Any suggestions?

sql – Can I (in a many to many relationship) select only those id:s in column A that has a connection to all unique id:s in column B?

I need to retrieve only those id’s in “A” that has a connection to all unique id´s in “B”.

In the example below, the middle row would be excluded because “A” does NOT have a reference to all unique id´s in “B”.

enter image description here

Is there a way to accomplish this in sql server?

I can´t really come up with a good discription/search term of what it is I´m trying to do (“Exclude column A based on values is column B” is not quite right). Hence I´m striking out looking for resources.

Do connection and message coexist in CSP / Pi calculus?

Berkely Socket API has two different kinds of network sockets:

  • byte stream sockets: connection-oriented and message-less
  • datagram sockts: connection-less and message-oriented.

In Pi Calculus (and CSP), there is both channel and message.

  • Is it correct that a channel is a connection? Or what is the difference between them?
  • How do connection and message coexist and work together?

Thanks.

Server running on Windows 10 as a service, OWIN self-hosted with WebAPI endpoints making an SSL connection without configuring a private key

I have created a prototype application the runs on Windows 10 that communicates with a server (described in the title) running as a service on a different system and successfully got SSL working but I’m missing something because I never seemed to generate any keys. I only just self-taught myself how to use PowerShell to create the self-signed certificate (so bare with me) using New-SelfSignedCertificate but I have very little insight into why this worked at all. From my very limited understanding a private key on the server is absolutely required for SSL to work, and I never associated one with my certificate… but it’s working and I think I’m fooling myself.

All I did was create the certificate with New-SelfSignedCertificate -Subject "CN=My Server Name"
Then add the binding with netsh http add sslcert certhash=<the thumbprint> appid={the app id}
And add the reservation for the service with netsh http add urlacl url="https://+:<my port number>/" user='NT AUTHORITYLocalService'

This all works, but I’m not sure I understand how because I never configured a private key. I know it must need to be expanded upon somehow to make it more secure because every single tutorial out there talks about making the certificate trustworthy and assigning a private key to the cert… but I’ve done none of that and it still appears to work. Granted, the client is not a web browser and will never care about the authenticity of the server because of the context in which this particular application will run (It’s to connect scientific components together in a lab, and there’s almost no exposure outside the network).

What am I missing here? Did Windows 10 provide a private key anyway? How can it really be encrypted if I didn’t need to specify these things? I also never moved the certificate to a trusted store of any kind because it frankly didn’t seem necessary since the client is just a proprietary front-end application.

ssh connection refused on Debian 7

I cannot ssh to my Debian 7 server anymore. It happened after I tried to upgrade some packages on this old distro (including libc6…). Uptime is almost 2200 days! I do not even want to risk restarting it.

ssh -V returns OpenSSH_6.0p1 Debian-4+deb7u7, OpenSSL 1.0.1t 3 May 2016

I somehow still have an active ssh connection on Putty as a root so I can still run commands. Hopefully I will not lose it before I fix this massive issue!

nothing outputs when I run /etc/init.d/ssh restart, it just prompts a new line for new commands.

ps -f -p $(pgrep sshd) shows a pid and uid root

/usr/sbin/sshd does not exist

cat /etc/ssh/sshd_config shows nothing any different than it has always been. I run it on port 22000. No issues since 6 years until now. I now explicitly opened the port 22000 on iptables just to make sure but nothing changed.

sftp -oPort=22000 -vvv root@redacted_for_privacy shows

OpenSSH_6.0p1 Debian-4+deb7u7, OpenSSL 1.0.1t  3 May 2016
debug1: Reading configuration data /etc/ssh/ssh_config
debug1: /etc/ssh/ssh_config line 19: Applying options for *
debug2: ssh_connect: needpriv 0
debug1: Connecting to redacted_for_privacy (123.456.redacted_for_privacy.123) port 22000.
debug1: connect to address 123.456.redacted_for_privacy.123 port 22000: Connection refused
ssh: connect to host redacted_for_privacy port 22000: Connection refused
Couldn't read packet: Connection reset by peer

command netstat -tulpn | grep 22000 returns nothing, however it does return udp6 for 22

command ps aux | grep sshd returns :

root     12247  0.0  0.0  29920  1252 pts/0    T    Nov21   0:00 vim /etc/ssh/sshd_config
root     12248  0.0  0.0  29924  1308 pts/0    T    Nov21   0:00 vim /etc/ssh/sshd_config
root     25974  0.0  0.0   6312   788 pts/0    S+   01:09   0:00 grep sshd
root     31003  0.0  0.0  81232  3932 ?        Ss   Nov21   0:14 sshd: root@pts/0

vim is because i set PermitRootLogin yes (it had always been set at without-password). 31003 is sshd pid

I also tried to reinstall:
apt-get install openssh-server but it returned
openssh-server : Depends: openssh-client (= 1:6.0p1-4+deb7u4) but 1:6.0p1-4+deb7u7 is to be installed
and E: Unable to correct problems, you have held broken packages. which is probably very true considering the ugly tinkerings I tried in order to upgrade libc6!

tail -f /var/log/auth.log when I try to ssh via WinSCP does not show any new line appending. either trying on port 22 or custom 22000.

netstat -ntlp returns

Proto Recv-Q Send-Q Local Address           Foreign Address         State       PID/Program name
tcp        0      0 0.0.0.0:21              0.0.0.0:*               LISTEN      12431/proftpd: (acc
tcp        0      0 0.0.0.0:3129            0.0.0.0:*               LISTEN      4374/(squid)
tcp        0      0 0.0.0.0:443             0.0.0.0:*               LISTEN      4917/nginx
tcp        0      0 127.0.0.1:3306          0.0.0.0:*               LISTEN      7231/mysqld
tcp        0      0 0.0.0.0:10000           0.0.0.0:*               LISTEN      10330/perl
tcp        0      0 0.0.0.0:80              0.0.0.0:*               LISTEN      4917/nginx

I tried behind a different ip, I’m sure I did not jail myself.

The 4 websites hosted on it currently keep working as if nothing ever happened.

I am out of ideas and desperate. What else should I check/try? I believe it is not running sshd but I have not idea how to restart it. I do not have systemctl, neihter service sshd restart, and nothing in /etc/init.d/