rvassar
Guru
- Joined
- May 2, 2018
- Messages
- 972
So I spent quite a bit of time hacking on my TrueNAS box this morning. I had been running an Intel i3-2120 for the last several years. Mostly because it supported ECC in my X9SCL, and was cheap/free/"available in my junk drawer". It got the job done, but didn't always keep up with Zoneminder frame analysis, or keep the 10GbE NIC fed. So I sourced a E3-1240v2, which should provide roughly 3x the CPU capability, and set about replacing it this morning.
Caveats encountered:
1. X9SCL BIOS 2.3 doesn't like 2021. I started a separate thread on this.
2. Initially came up in a panic reboot loop. It panic'ed right after importing the third pool, and printing the bit about swap using software crypto. I poked around at it for several minutes, and decided it might be corruption on the swap partition somehow. The last pool to import was the pool dedicated to my security cameras, and has become something of a dumping ground for drives of questionable reliability. This includes a 3Tb Toshiba N300 that @Chris Moore and I have discussed several times in the past. Most of the time it behaves itself and just works... Until it doesn't. I decided to pull the SATA cable off of it and see what would transpire, and sure enough the system came up. This appears to have forced TrueNAS to rework the swap spaces. I plugged the cable back in hot, and let it resilver, and it's been working correctly ever since, including several reboots. I'm not sure what the root cause of this actually was. I had not enabled the AES-NI instruction set, and the swap was using software encryption. After I worked thru the BIOS update, I did switch on the AES-NI option, and the boot process now reports hardware crypto in use.
3. This system was upgraded from 11.x, has been thru the iocage jail migration, etc... It is currently running TrueNAS 12.0-U3.1. It looks like I've accumulated some cruft along the way. I'm seeing the following in the boot process:
This pops in right after my APC UPS enumerates on the USB bus. I found some threads from last summer that seem related, but I don't seem to be experiencing any failure like those threads imply. Just kind of curious if I need to address this or not.
Results: I'm just getting back to the point where I have all my services & shares back in place. So far I seem to be running 8/C cooler, and my Zoneminder zone alarms are actually getting asserted in the frames where the events are actually occurring. The UPS power meter suggests I'm using perhaps 20 watts less power. This might be a result of the CPU being able to actually hibernate some cores.
Caveats encountered:
1. X9SCL BIOS 2.3 doesn't like 2021. I started a separate thread on this.
2. Initially came up in a panic reboot loop. It panic'ed right after importing the third pool, and printing the bit about swap using software crypto. I poked around at it for several minutes, and decided it might be corruption on the swap partition somehow. The last pool to import was the pool dedicated to my security cameras, and has become something of a dumping ground for drives of questionable reliability. This includes a 3Tb Toshiba N300 that @Chris Moore and I have discussed several times in the past. Most of the time it behaves itself and just works... Until it doesn't. I decided to pull the SATA cable off of it and see what would transpire, and sure enough the system came up. This appears to have forced TrueNAS to rework the swap spaces. I plugged the cable back in hot, and let it resilver, and it's been working correctly ever since, including several reboots. I'm not sure what the root cause of this actually was. I had not enabled the AES-NI instruction set, and the swap was using software encryption. After I worked thru the BIOS update, I did switch on the AES-NI option, and the boot process now reports hardware crypto in use.
3. This system was upgraded from 11.x, has been thru the iocage jail migration, etc... It is currently running TrueNAS 12.0-U3.1. It looks like I've accumulated some cruft along the way. I'm seeing the following in the boot process:
Code:
interface uhid.1 already present in the KLD 'kernel'! linker_load_file: /boot/kernel/uhid.ko - unsupported file type interface wmt.1 already present in the KLD 'kernel'! linker_load_file: /boot/kernel/wmt.ko - unsupported file type
This pops in right after my APC UPS enumerates on the USB bus. I found some threads from last summer that seem related, but I don't seem to be experiencing any failure like those threads imply. Just kind of curious if I need to address this or not.
Results: I'm just getting back to the point where I have all my services & shares back in place. So far I seem to be running 8/C cooler, and my Zoneminder zone alarms are actually getting asserted in the frames where the events are actually occurring. The UPS power meter suggests I'm using perhaps 20 watts less power. This might be a result of the CPU being able to actually hibernate some cores.
