Problems connecting to anonymous cvs servers

[jjstorm]

I attempted to connect, however, depending on the server, I was prompted for a password or permission was denied. I manually verified the fingerprint to make sure it is correct first. I also opened ports 21 and 22, incoming and outgoing on tcp and udp. Finally, i opened all outgoing traffic.

Since I have 5.6 release installed, I am following -patch branch. anything to check with the ssh client or ufw firewall?

By the way, I am running openbsd in a virtual environment.

thanks again.

[jggimi]

Based solely on the information provided, I can guess that something is wrong with your $CVSROOT (or cvs -d string).

It should be: anoncvs@<mirror>:/cvs

Review the examples in OpenBSD AnonCVS.

If any part of the string is incomplete or missing, your connection will fail.

[jjstorm]

Even though the string appeared to be entered correctly, for some reason, it did not work. When I entered it a second time, it accepted it. The process of downloading the source tree took over an hour. The string below (for the patch branch) worked. You can stick whatever server you want in there. Make sure to verify the server's fingerprint. Make sure you replace /amd64/ with your platform, and the tag to match your release.

# cd /usr 
# cvs -qd anoncvs@anoncvs.ca.openbsd.org:/cvs get -rOPENBSD_5_6 -P src 

then I preceded to rebuilding the kernel with the code below.

# cd /usr/src/sys/arch/amd64/conf
# /usr/sbin/config GENERIC
# cd /usr/src/sys/arch/amd64/compile/GENERIC
# make clean && make

I rebooted the kernel with the code below.

# cd /usr/src/sys/arch/amd64/compile/GENERIC
# make install         
# reboot

and finally I rebuilt the binaries which took about an hour and a half at least.

# rm -rf /usr/obj/*
# cd /usr/src
# make obj
# cd /usr/src/etc && env DESTDIR=/ make distrib-dirs
# cd /usr/src
# make build

To update the source tree going forward.

# cd /usr/src
# cvs -q up -rOPENBSD_5_6 -Pd

I am not fully clear on how updating the source tree, after the initial download of the entire source works exactly. Once I update my existing source tree with the above code, then how do I actually build the updated part of the source tree?

Also, going from release to -stable and subsequently updating -stable, do I ever need to update Xenocara?

Thanks again.

[jggimi]

Originally Posted by jjstorm

The process of downloading the source tree took over an hour.

Yes, the initial download of a working directory from a repository is not very efficient. This is why FAQ 5.3.3 recommends pre-loading source from a -release tarball.

I am not fully clear on how updating the source tree, after the initial download of the entire source works exactly. Once I update my existing source tree with the above code, then how do I actually build the updated part of the source tree?

Your personal /usr/src tree is -stable from the time you obtained it from the repository. It's known as a working directory. Each directory in that tree includes a CVS/ directory, with information such as the -stable tag you used (OPENBSD_5_6) and the repository you connected to. When you run a cvs(1) update command, the cvs client will review every directory and apply any new -stable updates to your working directory. Then, you would rebuild your kernel(s) and userland once again. If you have multiple -stable systems of the same architecture, you would build a -stable "release", and deploy by upgrade or install as needed.

...do I ever need to update Xenocara?

If there are any -stable patches for it, yes. If not, no.

Don't forget the ports tree. There can be -stable ports, too.

[ibara]

Originally Posted by jjstorm

Make sure you replace /amd64/ with your platform

s/amd64/`machine`/

[ocicat]

Originally Posted by jjstorm

The process of downloading the source tree took over an hour.

I can vouch that downloading from ca.openbsd.org is slow, & this is why the farm of anonymous CVS servers exists -- to divert traffic away from the master. It will be worth your time to study the list of servers --

http://www.openbsd.org/anoncvs.html

...to find an AnonCVS server which is geographically close to you. traceroute(8) can be useful to show network latencies.

The string below (for the patch branch) worked.
...

# cd /usr
# cvs -qd anoncvs@anoncvs.ca.openbsd.org:/cvs get -P src

If this is the command which was used, you did not download OpenBSD's -stable branch. As described in Section 5.3.3 of the FAQ, the correct options needed to download the -stable branch is:

# cd /usr
# export CVSROOT=anoncvs@anoncvs.example.org:/cvs
# cvs -d$CVSROOT checkout -rOPENBSD_5_6 -P src

Without explicitly providing the tag OPENBSD_5_6, source at the head of CVS, which is called -current, will be downloaded. If -release packages are installed on top of -current, you will most likely experience mismatches which are not supported. More information which needs to be understood when building OpenBSD is found at:

• Section 5.1 -- describes how OpenBSD organizes its CVS tree.
• Section 15.4 -- provides further information on the unsupported practice of mismatching the installed system with packages/ports.

[jjstorm]

Originally Posted by ocicat

I can vouch that downloading from ca.openbsd.org is slow, & this is why the farm of anonymous CVS servers exists -- to divert traffic away from the master. It will be worth your time to study the list of servers --

http://www.openbsd.org/anoncvs.html

...to find an AnonCVS server which is geographically close to you. traceroute(8) can be useful to show network latencies.

If this is the command which was used, you did not download OpenBSD's -stable branch. As described in Section 5.3.3 of the FAQ, the correct options needed to download the -stable branch is:

# cd /usr
# export CVSROOT=anoncvs@anoncvs.example.org:/cvs
# cvs -d$CVSROOT checkout -rOPENBSD_5_6 -P src

Without explicitly providing the tag OPENBSD_5_6, source at the head of CVS, which is called -current, will be downloaded. If -release packages are installed on top of -current, you will most likely experience mismatches which are not supported. More information which needs to be understood when building OpenBSD is found at:

• Section 5.1 -- describes how OpenBSD organizes its CVS tree.
• Section 15.4 -- provides further information on the unsupported practice of mismatching the installed system with packages/ports.

I know about the server list, and I was aware of tracerout. I just did not think to use it.

Indeed I did install the -stable source tree as my installation of openbsd shows during the boot process. You caught a paste error on my part that I am fixing as we speak, although I entered the string slightly different. see above.

I have read all of the faq, however, I have not been able to grasp all of the concepts as I am a total noob to unix. I have also read another popular OpenBSD book that was published recently. The reading that I have done has been exhaustive, but I know that is what it takes especially for a noob. I will keep at it. Thank-You very much for your input.

[jggimi]

Originally Posted by jjstorm

...I have not been able to grasp all of the concepts as I am a total noob to unix.

It just takes time, and some trials, and some errors. I think you're doing extremely well, by the way.

I have also read another popular OpenBSD book that was published recently.

Ah, no wonder you're doing so well! I keep both editions of Absolute OpenBSD on my shelf, side by side. (And not just because I have a couple of my Haiku published in the 2nd edition. There was a contest.) It's a great reference!

Mr. Lucas has an account on this forum, but hasn't logged in for a long time. He's active on Twitter and posts regularly on his blog.

[jjstorm]

Originally Posted by jggimi

Ah, no wonder you're doing so well! I keep both editions of Absolute OpenBSD on my shelf, side by side.

Actually, the book is how I came across daemonforums, and thanks for the compliment.

[shep]

Although learning produce your own -stable OpenBSD system is a good learning exercise, it is time and resource intensive. You can arrive at the same endpoint by utilizing M:tier. M:tier is a recent offering by a prominent contributor to OpenBSD, Antoine Jacoutot. This option became available after the release of Mr. Lucas's 2nd edition. You can entirely automate the process using # openup or just use the port updates. I do my own patching and rebuiliding of the base install but utilize M:tier package updates.

[jjstorm]

Originally Posted by shep

Although learning produce your own -stable OpenBSD system is a good learning exercise, it is time and resource intensive. You can arrive at the same endpoint by utilizing M:tier. M:tier is a recent offering by a prominent contributor to OpenBSD, Antoine Jacoutot. This option became available after the release of Mr. Lucas's 2nd edition. You can entirely automate the process using # openup or just use the port updates. I do my own patching and rebuiliding of the base install but utilize M:tier package updates.

Like you reference , for now, I would prefer to do as much manual labor as possible to get a better understanding of the system. looks very interesting by the way. I bookmarked the site.

[jjstorm]

My current situation is a little bit different than when I started this thread.

I am currently running 5.8 release, and I already have manually downloaded and unpacked the source files (src.tar.gz and sys.tar.gz) to /usr/scr

The following example can be found under anoncvs:

NOTE: If you are updating a source tree that you initially fetched from a different server, or from a CD, you must add the -d [cvsroot] option to cvs.

# cd /usr/src
# cvs -d anoncvs@anoncvs.ca.openbsd.org:/cvs -q up -Pd

Although they don't confirm it, the above example appears to be for following -current. How would I change this for following -stable keeping in mind that I manually unpacked all the source files?

[jggimi]

As discussed in 2014 in this thread, please use a repository closer to you.

And to ensure you are obtaining 5.8-stable, use -r OPENBSD_5_8 with the update command.