I started trying a Root-on-ZFS upgrade using beadm and quickly encountered an anomaly: no /usr/bin Then, it was there, then it wasn’t. I wasn’t sure of what steps I took to reproduce it. So I went through logs and eventually understood what happened. Let me show you. Here is what I did: # beadm create 10.1-RELEASE # beadm mount 10.1-RELEASE /mnt # mount -t devfs devfs /mnt/dev # chroot /mnt csh: Cannot open […]
now it’s not there, now it is Read More »
This is tallboy, the server I have at RootBSD. I like to store away this type of documentation, in case I need it later. This server was originally on UFS with gmirror. The partitions: The ada0 drive details: And ada1 drive details: And finally, dmesg:
I just converted another server from UFS to ZFS. I used this script as my starting point. I dropped to a shell as soon as I booted from the live CD. Then I fetch my script: fetch –no-verify-peer http://bit.ly/1MvhRXH mv 1MvhRXH install.sh I didn’t run the script. I modified it to suite my new conditions. The install files were not available. I put and exit around line 84. Then I ran my script:
Another zroot install Read More »
In this post, I am using FreeBSD 9.3 but what I do here should have wide application, especially the errors I encounter. I believe in backups. Not just rsync, but multiple copies and in multiple places. RAID is not a backup. rsync is not a backup. The goal of this post is to send a copy of a ZFS filesystem to a remote server. The server in question is already backed up in
zfs send | zfs receive as non-root Read More »
Full disclosure: I’m biased. I like the combination of PostgreSQL, FreeBSD, ZFS, and Bacula. In fact, I like it so much, I wrote the PostgreSQL backend for Bacula. Since then, I have been using Bacula for over 10 years. I have backed up to HDD, DDS, DLT, and SDLT. Bacula has proven to be reliable and it just works. Clients are available for a wide range of operating systems and Bacula doesn’t care
Bacula on FreeBSD with ZFS Read More »
Here is the output for zfs-stats on slocum. This represents a 24 hour period. This server acts as my database server (MySQL and PostgreSQL), and runs bacula-dir, but does not store any backups locally. It also houses the dev copy of FreshPorts.
zfs-stats for slocum Read More »
Here is the output for zfs-stats on knew. This represents a 24 hour period. This server receives backups (via bacula-sd) and runs regression testing in various jails for the Bacula project. As a little side venture, it also loads up each of my backed-up databases.
zfs-stats for knew Read More »
NOTE: when upgrading a zpool, it may be necessary to also upgrade the bootcode for that disk (or disks). Be aware of this before proceeding. I know I have missed this step more at least once. Earlier today I was adding a new pool to the system. After creating the pool, the status showed this: Let’s upgrade that pool now Upgrading the pool Upgrading the pool is easy: Oh… Now what? I wasn’t
zpool upgrade calls for gpart bootcode -b /boot/pmbr -p /boot/gptzfsboot Read More »
I haven’t tested this yet, but I’ve installed the hardware. The real test will come after a reboot and actual use of the new drives. NOTE: Although this claims to be creating a new ZIL (SLOG) do not believe it. In fact, stop reading right here. There is nothing of use below. What will they be used for? These drives are intended for ZIL (the ZFS Intent Log). This can speed up writes.
Adding two new drives (SSD) to an existing ZFS system Read More »
It seems that when I decided to send a filesystem from one server to another, I neglected to establish sufficient space existed. This morning, before I headed to BSDCan, I found that my server was very sluggish and slow to respond. Nagios was flagging all kinds of errors, some of which I’d never seen before. Looking at the system in question, I found the system 95% full: $ zpool list NAME SIZE ALLOC
zfs receive stalled, filesystem 95% full Read More »