beastie is up again

beastie is finally up

beastie is a new accessible host inside sdf clusters which uses OpenBSD as the OS and same $HOME directory as other sdf's hosts.

While beastie is great, frequent fsync errors may occur randomly because there are chances that NFS mount shared between sdf clusters is being non-responsive. The last of this error causes several days downtime on beastie host.

The last time I encountered the problem on beastie is several days ago. During the downtime period, users are unable to establish new ssh sessions to beastie for several days.

One host is down, but the cluster is accessible

Fortunately, sdf has several hosts that can be used to log in. The available hosts can be accessed via tty.sdf.org and each provides access to the same $HOME, so one's inability to access beastie host is actually no problem at all.

As of today

beastie is up and running. I am impressed with lightweight memory use of beastie, several days uptime with very minimal memory usage. Probably beastie doesn't run as many service as other tty hosts.

      _____       mydeardiary@beastie
    \-     -/     -------------------
 \_/         \    OS: OpenBSD 7.0 amd64
 |        O O |   Host: Supermicro X7DBT
 |_  <   )  3 )   Uptime: 10 days, 33 mins
 / \         /    Packages: 192 (pkg_info)
    /-_____-\     Shell: mksh R59 2020/10/31
                  Terminal: /dev/ttypo
                  CPU: Intel Xeon X5460 (8/8, SMT off)
                  Memory: 198MiB / 16367MiB

So as my preferred host on sdf is up, I can get back enjoying sdf.org as $HOME, far away from home.

If you are fellow sdf user, give beastie a try by logging in to beastie.sdf.org. Remember to set up ~/.ssh/authorized_keys beforehand to make logging in easier.

One useful tip: be patient

It seems that one needs some patience when one of the host on the cluster is not accessible.

Well, done. Long life sdf.org as the oldest public access system.