superloopy1

Thanks but it wont help. Unbalance seems to work at the disk level when reorganising data. What i want is something which will take a share as input and consolidate its data correctly across a range of disks. Not sure Unbalance works that way. Could be wrong though.

Just received a few of these error messages in syslog? "nginx: 2023/05/26 16:21:57 [error] 7470#7470: nchan: A message from the past has just been published. Unless the system time has been adjusted, this should never happen." Any ideas? System is currently running Mover, might be important?

My latest system has got itself in a mess when it comes to what media, plex mainly, is where and i need to reorganise everything back on single disks for ease of access and backup consistency. Whats the quickest way to do this on lets say a tv show which has propagated itself across many disks due to, i assume, the wrong split level on the share. I need to tidy up ..... i cant be the first to have to do this across their disks.

Happened again ... i'm now on 6.12 rc6 lets see eh.

Thanks JorgeB but just too late unfortunately. It hung up altgether and needed a hard reset. Next time eh!

Came back to my PC after a few hours and for the n'th recently time the webgui was again inaccessible bar the top Unraid banner. Only tool i could think of was ssh so in i went, quick 'top' later and this is what i see ,,, 344 root 20 0 0 0 0 R 100.0 0.0 134:17.62 kcompactd0 1640 root 20 0 4232 2384 2072 R 100.0 0.0 35:33.28 pgrep 1892 root 20 0 4364 2600 2016 R 100.0 0.0 35:33.29 pkill 11147 root 20 0 0 0 0 R 100.0 0.0 36:05.60 kworker/10:0+rcu_par_gp 13769 root 20 0 4232 2336 2012 R 100.0 0.0 35:33.29 pgrep 15144 root 20 0 4232 2336 2024 R 100.0 0.0 35:33.29 pgrep 50671 root 20 0 4232 2292 1972 R 100.0 0.0 129:47.71 pgrep 50908 root 20 0 4232 2296 1976 R 100.0 0.0 129:27.74 pgrep 55244 root 20 0 4168 1136 984 R 100.0 0.0 35:33.30 ps 12442 root 20 0 4232 2336 2020 R 100.0 0.0 35:33.29 pgrep 13379 root 20 0 4232 2264 1940 R 100.0 0.0 35:33.29 pgrep 50515 root 20 0 3836 2392 2004 R 100.0 0.0 129:58.44 sensors 65421 root 20 0 4232 2328 2012 R 100.0 0.0 35:33.29 pgrep Just repeating on these processes, everything below is at next to no CPU usage, obvious really. System IS still up and running at 100% cpu but how do i recover this without just hitting the buttopn, which is how i would normally have to do to fix this. Is there any way of stopping those 'processes' cleanly and restoring a usable system? I'll just leave things for now, see how any responses go (but electricity is expensive here in the UK ;-)))

aarghh. thats the way i had it originally before it stopped delivering the goods. i'll change it back when i get a minute. i've also changed a lot of SMB and UD settings in my attempts to get this working. Guessing i'll have to rgress these at some point if i can remember what they were set to ie defaults. Thanks again!

Thank Thanks guys, i am now sorted, and yes, it wasset up originally as you lay out Squid. Problem was of my own making (possibly). When the share last worked it was set up using the IP address 192.168.1.5, when i recreated it i used the server name xxTOWER. Seems the mismatch doesnt resolve very well and didnt match the docker paths which were still pointing at the IP address. I'm still unsure as to which is most efficient, IP or servername? Once corrected it place a xxTOWER share on the receiving machine and after a scan all folders are now showing. Problem is it WAS working perfectly then it didnt and I never went anywhere near the configs hence the reason for forgetting what was in place when i deleted and attempted to reset it. Honest guv ;>) it wasnt me !

Thanks. yes I did but i'm not sure i did it correctly. is it the case that for server B to access server As files then the share needs setup on BOTH servers or just SERVER B? and, ive moreorless activated all smb settings in UD settings and still my shares are not populating on either emby or plex whereas they used to do so no problems. Which diagnostics do you need?

I used to have a couple of SMB network shares between my two servers but recently they stopped working and i deleted them. No i have no idea how i set them up in the first place, permissions etc and cant seem to 'see' the servers at all, d'oh! If i want to shae from server1 to server2 which machine gets the 'share' set up and as permissions seemed to be my downfall last time around what's needed here? Just need to see the shares on the other machine. Seem to remember i set it up from the main page thats all . Cheers ...

Thanks for that. Unassign is a simple as that eh, It can all be done without a powerdown?

I should know this but its solong since i touched this particular array and i need someone to explain the procedure. Disk1 eg is almost full with no prospect of relocating the contents elsewhere. I have no spare disk slots but DO have a larger drive withing the capacity of the overall parity. Is it a simple drag the old drive out and swap the new drive in or is there some unassign process needed? My minds just not remembering the 'good' procedure any more ... thanks guys.

Well now. I'm on 6.5 and can no longer 'see' the plugin in the appstore. Has it been pulled from the production ie stable release line?

I got fed up with seeing the prompt that disklocation app could not be loaded against my 6.11.5 system until i had 6.11.9 at least. Surely this prompt should not be being shown on a system running the most stable release 6.11.5 when i check for app updates. So ... i just removed the app, too much of a pita. Suggest others who find this a distraction ie 'when is an update not an update', answer being when its being offerref up in error and not corrected, just delete the app.

Ah, so there's the catch then. I'll need to raid10 the ssds outside of the array which to all intents and purposes is just a front to enable my dockers to run. Trim CAN be used against those cache pools external to the array. I'm learning i think .... thanks again, very prompt!

Is TRIM available within the array if parity is NOT being used?

That sounds promising. So if i exclude in global share settings all activity will only be to cache pool shares and will run at the speed of the slowest ssd in the system? No parity drive needed anywhere? Really need to understand cache pools now before i drop an awful lot of cash on hardware. Thanks!

Many thanks, i'll continue on then. Mobo supports 4 x nvme2 and 8 sata so i'll try and populate these out as best possible then install unraid and dockers.

Presumably if not being written to then trim wouldnt be necessary? Any way to trim the pooled ssds? The more i learn the less convinced i am that this is a good idea ...

Could a 'gash' ssd serving no useful purpose be used IN the array to satisfy this requirement and if so would it need to be comparable in speed to the pool ssds. I'm reading a lot about unraid operating at only the speed of the lowest drive in the array? Is this only when you ARE actually writing to data disks themselves. In my case i'll be writing to pooled cached drives, probably cache10. Seems a bit messy in order to run up a working Unraid system when just wanting to get as much processing control and minimal power out of the hardware.

The ssd pool will be short term storage, once processed, hopefully a lot quicker thn at present, then it will be moved off to one of three longer term storage arrays. So this small, efficient nas is only there as a frontend. The data WILL be held on the ssd pool but not long term. Cronjob will ship off on a regular interval. I gather this on/off writes to the ssd pool is the reason why unraid wont allow to be used in a full array as spinning disks are. My unraid array wlll therefore only be a couple of cache pools of ssds with no spinners to be seen. NVME slots are available but i'm guesing they'll be for metadata. once working, its all to be shipped off via 10Gb/e to a final resting place. I'm sure this is being used by others but cant find a decent writeup of the whole process .... thanks for the help!

Wow, quick, thanks! Can i run this past you then ... 8 disk nas, all ssds, how do i configure an Unraid system? No parity i assume, not needed as all content to be shipped asap to a 'suitable' array for storage. This lean nas is purely to do all the donkey work and allow the running of dockers, within Unraid? all post processing, unpacking etc to be done and then short term stored on an ssd structured cache pool, yes? Am i understanding or not? Would 10GBe be of use in shipping out no longer required filestore to the slower spinning disk array? Tell me if i'm talking bilge here ....

I'm trying to plan out a small Unraid system which is both powerful and efficient. Initailly i wanted a small nas utilising SSDs but since reading that ssds are not recommended to be IN an unraid array i'm wondering what the next best option is? I already have three storage arrays which i intended to use to service this smaller, leaner array which was going to be mini-itx with all ssds/nvme m2 from the get go. This was to be the workhorse frontend to the 'arrs, plex, dockers and further down the line VMs as i understand whats best needed. But now, without ssds as that backbone what are my options. This smaller nas was to have 10Gb/e (just because it can and as possible future proofing) coupled with a 13th gen intel cpu utilising quicksync capabilities. A lot of this is MY wishlist and, i understand, not necessary in order to transcode plex/emby. I understand this. Now, left without an ssd based objective, how do i proceed? Maybe i just stick with the xeon based storage i have, maybe consolodate the lot into one larger case and forget the 'efficiency' angle altogether. I cant help thinking this is not the way i saw my use of ssds progressing in the longer term. Will Unraid ever sanction an ssd based array i wonder and If not then what are others using as a good alternative to run a 'lean mean Unraid' system.

Thanks guys, couple of suggestions there that i could take up. One non-runner is moving my server to take advantage of the Xeon (mega!) haeat output, my wife would kill me if either of my three servers ever became visible around the house , that's why i need to go with a mini-itx based system and only a couple of nvme/ssd drives so an almost silent system. I DO have electricity at slightly higher rate than yours across there overnight so i'm thinking along similar lines and 'moving' lets say a weeks worth of downloads, again everything now has to be on cheap rate electricity, via rsync nightly from a cached pool on my main servers to the small, silent itx system in my lounge. Might work? But, it means i cannot then access any back catalogue stuff, a big downer on the 'wife' front. Seems it cant be done in a 'wife-friendly' manner ... so, it stays as it is, zilch downloads.

Can i just vent my spleen over electricity prices here in the UK and ask other forum members for some constructive advice. Power costs are horrendous here in the UK (and no doubt elsewhere) to the extent that i cannot financially afford to run up my Unraid servers for any length of time. Unfortunately i have three Xeon based servers consisting of some 30+disks. Main use is storage and obviously, like others streaming as a nas to my own local devices. This has had to stop as has adding any new content to my machines. Its a pretty sad situation which i'm looking for ideas on how to minimise powering up the lot in one go and still have an effective subset of my main NAS available in a leaner, meaner form. The drive sizes range from 10TB to 18TB so obviously rolling up to larger drives is really not an option as it would mean shelling out for 20TB drives which just isnt do-able. Just wondering how i can tweak my setup to just have a 'small' nas setup of some 1-2 drives and how its possible to offload any files wanted 'automatically' until watched? I do have cheap rate (sic) overnight electricity but whats a decent, cheap setup to allow up to 4K playback, whether direct or transcoded? Best bare minimum needed anyone? All suggestions gratefully received. Ta!