In this power, the server is running FreeBSD 10.3. I am completely and unreasonably biased: ZFS is the best filesystem out there. Do not take my word for it. Ask around. Today, I started a process I’ve been waiting to do for a while. I am replacing the 3TB drives in a 10-drive raidz2 array with 5TB drives. These new drives are faster, I think perhaps cooler, and, more to the point, bigger. […]
Swapping 5TB in, 3TB out Read More »
This is mostly a note to myself. #### Edit: 2018.09.08 : If you see messages like this: $ sudo gmirror insert swap /dev/da5p2 gmirror: Class not found $ gmirror list -a gmirror: Command ‘list’ not available. You probably need to do this: sudo kldload geom_mirror #### I had these drives: They are involved in this zpool: I decided to create a 6-way mirror for swap: sudo gmirror label -v -b round-robin swap diskid/DISK-653DK7WBFS9Ap2
Creating a gmirror swap Read More »
Last night, when I got my FreeBSD & ZFS based Time Capsule running, I had to connect my laptop to the same network as the server in question. This is not ideal. My usual work flow: connect to the WIFI, then connect to the VPN, then I get access to those services. These are different subnets, so the Netatalk broadcast does not traverse the switch. It stays inside so my laptop does not
Accessing your Time Capsule when on a different subnet Read More »
I have a system with three tape drives and two tape changers. If one tape library is powered off when the system boots, the device names for the other tape library may be skewed. That is, /dev/sa0 may not be the LTO-4 drive, it will be the SDLT drive. This is not ideal. FreeBSD uses device.hints for this. I have used it before, and for quite some time, however, I learned something new
Using device.hints to wire physical devices to specific names Read More »
The knew server is ‘frozen’ again. This has been happening daily at about O301 UTC each night. See my Twitter feed for background. In this post I will include details as I progress through the data. The server in question is knew (yes, that’s the hostname). dtrace hotkernel I left this running in an ssh session and pressed control-C this morning: [root@knew:/usr/share/dtrace/toolkit] # ./hotkernel >> /var/tmp/hotkernel ^C ssh login loop It was suggested
Server freeze – 2014.12.14 Read More »
I recently added 10 new HDD to a system which already had 10 HDD (now a total of 20). The HDD are split into two zpools: $ zpool list NAME SIZE ALLOC FREE EXPANDSZ FRAG CAP DEDUP HEALTH ALTROOT system 27T 22.1T 4.93T – 19% 81% 1.00x ONLINE – tank_data 45.2T 26.3T 19.0T – 37% 58% 1.00x ONLINE – zoomzoom 436G 9.67G 426G – 6% 2% 1.00x ONLINE – The zoomzoom zpool is
system freezes up with lots of sonewconn Listen queue overflow Read More »
WARNING: This did not work. It succeeded, without error, but the card did not work. There will be a new blog post soon. Yesterday, a new LSI SAS 9201-16i arrived. I bought it on eBay from a supplier in China and paid for expedited shipping. I offered US$250 for their $338 listing and it was accepted. Shipping was another $30. It arrived 6 days later. Why do I need one? The system in
Flashing an LSI SAS 9201-16i Read More »
After upgrading all my servers to FreeBSD 10.3 last weekend, I started to see these messages in the logs: Aug 3 21:15:19 slocum ntpd[29456]: leapsecond file (‘/var/db/ntpd.leap-seconds.list’): expired less than 64 days ago I found this post which suggested running service ntpd fetch but that did not fix it. While debugging the issue, I added ntp_leapfile_fetch_verbose=”YES” to /etc/rc.conf, which provides more output to the fetch command: $ sudo service ntpd fetch ntp_src_leapfile version
leapsecond file (‘/var/db/ntpd.leap-seconds.list’): expired less than N days ago Read More »
Let’s try the speed test which comes with Bacula’s btape utility. This is Bacula 7.4.0 on FreeBSD 10.2, with a Dell TL4000 LTO-4 tape library.
Using Bacula’s btape speed test on FreeBSD 10.2 and an LTO-4 tape drive Read More »
I have recently added two 480 GB SSDs to a 10 x HDD raidz2 system. The SSDs will be mainly used for spooling to tape during backups, but I’m going to use a small part of it for a SLOG. Not all systems benefits from a SLOG (Separate intent LOG), but synchronous writes, such as databases, do. All ZFS systems have a ZIL (ZFS Intent Log); it is usually part of the zpool.
Adding SLOG to a zpool Read More »