Zoneminder filing up drive quicker than I can delete events

[zerocool]

I have recently moved from using zone minder on Ubuntu for many years and moved over to Truenas as it ticked a lot of my other requirements but I have an issue where zone minder is filling the 8 TB drive set aside for its own use quicker than the purge filter can delete. I am having to manually delete events and stop zone minder for days to catch back up. as a rough guide if I delete large chunks of data in Truenas shell it deletes around 60 gig in 60 minutes so takes days to delete folders of one cameras worth of events. so im ordering if this is normal trueness speed or if I have something set up wrong. I have an 8TB purple drive which is set as 1 pool and mounted for the zone minder plugin to use for the events folder to be stored on. Also asked it to scrub and its taken 24 hours to get to 9%.

Hope someone can suggest something to try as I have read hundreds of posts and still can't fix it.

 

[Redcoat]

Hope someone can suggest something to try as I have read hundreds of posts and still can't fix it.

Welcome to the forums!

To get help here for a question like yours, at the least you need to provide full details on your hardware and FreeNAS/Trueness version - see the "Forum Rules" on the masthead if you're in doubt - the list there includes:

• Motherboard make and model
• CPU make and model
• RAM quantity
• Hard drives, quantity, model numbers, and RAID configuration, including boot drives
• Hard disk controllers
• Network cards

 

[zerocool]

HI Redcoat and thank you for the reply and the is my first ever forum post as I don't like to bother people but I'm stuck.

TrueNAS version TrueNAS-12.0-U2.1 and the zone minder jail is RELEASE 12.1 and ZM version V1.34.21
Motherboard is an MSI tomahawk
CPU is AMD Ryzen 9 3900x 12 core with AMD cooler
Ram 64 gig in 16 gig sticks corsair
3 8 TB WD red drives in raidz1 and 1 8TB WD purple just as a single drive used only for zone minder events. Boot drive is a crucial 120 GB SSD
Hard disk controller is LSI 92118i with firmware changed to version in TrueNAS manual
Network card is gigabit intel E1G42ET - 82576 chip as suggested on TrueNAS forums

I have 3 Jails and 1 Ubuntu VM running.

Jails are Plex, Base with Asterisk installed , and transmission plugin and the VM is running ubuntu 20.04 with zabbix on as I couldn't get this running in a jail but the VM has 2 threads and only 2 gig of ram.

I feel the hardware is very capable and shouldn't be struggling. I have also turned off compression on the Events drive to see if that is causing my issue.

Let me know if anything else is useful to know :)

 

[Redcoat]

Take a look at this thread https://www.truenas.com/cohttps://www.truenas.com/community/threads/very-slow-directory-file-deletions-from-shell.87287/mmunity/threads/very-slow-directory-file-deletions-from-shell.87287/ , particularly post #13 - might be some clues/ideas there.

Also - how full is the purple drive when you begin a deletion exercise? Do you have plenty of space left? Have you run top while the delete is running to see what's consuming CPU?

@Stilez may be able to chip in here with some pointers...

 

[zerocool]

HI and I had read that post as I found it and have looked at the things it mentions but nothing really helped.

Im not sure if my lack of knowledge is my own issue but moving ZM form a ubuntu KVM and I kept the purple drive at 95% as I wanted to maximise how much history I had I set up the one on trueness the same and have since read it isn't good to use that much space with ZFS so now I have got it back down to 68% full and set ZM to delete at 70% so im going to see what its doing tomorrow as ZM creates around 900Gb every 24 hours and the scrub is still running and says its got 6 days left so don't want to do too many things all at the same time but will post my status tomorrow.

My CPU is at 12% in top and using 30Gb of ram for ZFS cache with 28 Gb used for services but this is mostly ZM as it drops when I stop that jail but this is a normal ZM thing and reason I put 64Gb in when I built this unit.

 

[Redcoat]

Let's hope that the increased space provides sufficient headroom to speed up the delete process.. From what I know (which isn't an awful lot about "pushing the limits" of ZFS performance) that should help.

 

[zerocool]

Hi Redcoat I thought I would let you know it seems to be holding happy at 70% so hopefully its fixed and the scrub has finished now too. Thank you for the help :)

 

[Redcoat]

Thank you very much for the feedback. I'm glad to hear that you have made progress.