What’s the process for this? I’ll have to do it at some point.
They throw up the steps in the wiki…
Usually just changing the repos… a couple commands… an update…then a reboot.
1 Like
Am I right about this?
8x vs 16x pcie for gpu transcoding…
1 Like
Realistically, you could get away with 4x 3.0.
Transcoding is absolutely processing limited.
Depending on the gpu, you might need 8x, but a p2000? 4x is plenty.
1 Like
Thanks, that’s what I figured.
Someone shared this on one of my other channels:
5 Likes
that was a fun read
4 Likes
Glad to hear. I’ll be reading that when I get some time (probably on the plane)
1 Like
Work or pto?
Work. Should be interesting though.
1 Like
It’s been a minute since my posts last week @NX211 @Dynamic_Gravity @SgtAwesomesauce
But here is how I approached the problem. The Galera cluster is using Mariadb. You can query a Mariadb using SQL to get it’s status and the status of the cluster. PHP handles this well. PHP also has real interfaces, so I just used two of the well known object oriented design patterns to put together a monitoring app (script, whatever) that can query each node and get it’s state. In addition, it’s extendable so that I can add other types of nodes in a different type of cluster (think rabbitMQ, etc), just by targeting the node or “subject” interface.
I’ll have to do the state handling next, but that’s where I’m at.
6 Likes
Nice!
A good implementation of PHP is infrequently seen these days, so that’s cool.
4 Likes
I’ll take “things you never thought you’d read” for $200, Alex.
2 Likes
You guys need to follow the news dump thread…
I’ve been getting email alerts all day that 2 of my Unifi switches at home are restarting every couple hours… they are POE-powered, so I suspect that may be the cause, but idk where to look for clues other than the event log on the Unifi Controller which just has a notice about the reboot…
InfoSux department is losing their shit because I made an empty S3 bucket public for 30 seconds whilst I ran a Jenkins job.
We’re not allowed to have IAM accounts or roles that aren’t created by our generous masters, and we’re limited to five. The job fails because of lack of permissions. I have a draft standing by but it’s going to cause already high tensions to grow between the departments, which is not teh DvOpz wei.
So, has anyone run into this issue? This whole least privilege shit pisses me off when it’s applied to systems engineering teams.
Hello fellow Sysadmins -
My Fedora VDI blew up this morning and I need a hex editor to try to repair it. The host machine is Windows 10. Can anyone recommend a hex editor?
Kind of. We kinda solved this problem by remaining platform agnostic and being able to docker-compose a nearly identical to production environment on dev machines.
Giving devs free reign over dev environments that don’t use any actual data helped a lot in the dev speed.
Moving from Fedora 29 -> 30. Giddyup!
2 Likes
Put my Unifi DC instead of POE, but it’s still rebooting itself every 30-60 min. Crash dump has a lot of this:
Summary
Task PID: 805
Task pthread: 0xb5daa520
Task Name: bcmMEM_SCAN.0
/lib/libpthread.so.0(__pthread_cond_timedwait+0x240) [0xb6ece794]
0xf3ccf0: ????? (0x0) + 0x0
0x11bb3f4: ????? (0x0) + 0x0
0x571834: ????? (0x0) + 0x0
------------
Task PID: 804
Task pthread: 0xb6051520
Task Name: socdmadesc.0
/lib/libpthread.so.0(__pthread_cond_wait+0x19c) [0xb6ecec3c]
0xf3cd60: ????? (0x0) + 0x0
0x11d9354: ????? (0x0) + 0x0
0x571834: ????? (0x0) + 0x0
------------
Task PID: 803
Task pthread: 0xb60c8520
Task Name: bcmINTR
------------
Task PID: 802
Task pthread: 0xb6773520
Task Name: bcmDPC
/lib/libpthread.so.0(__pthread_cond_wait+0x19c) [0xb6ecec3c]
0xf3cd60: ????? (0x0) + 0x0
0xf3b004: ????? (0x0) + 0x0
0x571834: ????? (0x0) + 0x0
------------
Task PID: 794
Task pthread: 0xb677d520
Task Name: osapiTimer
/lib/libpthread.so.0(__pthread_cond_timedwait+0x240) [0xb6ece794]
0x5743f0: ????? (0x0) + 0x0
0x571834: ????? (0x0) + 0x0
------------
Task PID: 793
Task pthread: 0xb6787520
Task Name: EmailAlert
/lib/libpthread.so.0(__pthread_cond_wait+0x19c) [0xb6ecec3c]
0x575788: ????? (0x0) + 0x0
0x567c90: ????? (0x0) + 0x0
0x159b4c: ????? (0x0) + 0x0
0x571834: ????? (0x0) + 0x0
------------
Task PID: 792
Task pthread: 0xb6791520
Task Name: procLOG
------------
Task PID: 791
Task pthread: 0xb679b520
Task Name: LOG
/lib/libpthread.so.0(__pthread_cond_wait+0x19c) [0xb6ecec3c]
0x575788: ????? (0x0) + 0x0
0x567c90: ????? (0x0) + 0x0
0x156540: ????? (0x0) + 0x0
0x571834: ????? (0x0) + 0x0