RealActorRob
-
Posts
129 -
Joined
-
Last visited
Content Type
Profiles
Forums
Downloads
Store
Gallery
Bug Reports
Documentation
Landing
Posts posted by RealActorRob
-
-
On 5/19/2022 at 1:17 PM, ljm42 said:
This should do it:
/etc/rc.d/rc.nginx reloadFYI - Unraid 6.10 is a bit more strict with making sure your certificate matches your settings. I'd recommend running the Fix Common Problems plugin, it will warn you if your certificate is actually valid or not before you upgrade.
Also worked for me, boom.
What things should I look for to stop from crashing tho?
-
Specifically:
modprobe –r acpi_pad
modprobe: FATAL: Module –r not found in directory /lib/modules/6.1.64-UnraidWorks:
modprobe --remove acpi_pad
modprobe -h says "-r, --remove"
So why is it not happy with what should be the same flags?
-
If anyone is running a storj node, it likes to leave open files even with docker stopped.
Use the "Open Files" plugin to see what it's not closing and kill that process after making sure the Docker is stopped. This happens fairly often.
In this case, CLI from web works if this plugin doesn't....window popped up blank so :
kill 41524
-
Yeah, I'm looking for a good price on a drive...might get a lightly used datacenter large drive then replace it with one of the 4TB parity drives.
Worst case if it fails I have a 4TB SMR or just swap them back...plus i can tolerate 2 fails.
Note @JorgeB that I also changed the CPU governeor to 'on demand' and it's a bit faster tho it wasn't being pushed before anyway.
-
@itimpi Disk 6 seems a little wonky but here's diskspeed, ran just now DURING the parity check. It's getting pretty old, at over 7 years but SMART values are all 0 not even a CRC error. Cool at 40C as well.
-
Still slow, but still don't see how they should be this slow.
Only one bench didn't have a steady curve from Diskspeed, one of the parities.
I watched it for a bit and never saw it peak over maybe 30 on one for a few seconds.
Here's total throughput and something weird on reads:
Obviously this isn't correct but maybe a clue?
And I don't see how disk speed benches ok and then the parity check crawls, haven't had a fast one since March.
And that's across the board slow not just sections of some disk.
Real head scratcher.
-
No errors, drives bench way fast, total drive throughput is about 35-40 max.
11 days projected to complete, largest drive 4tb. No SMR.
For some reason two of my drives are slow at 50ish each (will address that) but there's no reason I can find for a sub 4Mb/s parity check.
Controller links at 4GB/s.
Diags attached.
-
12 hours ago, ljm42 said:
Jumbo frames (especially on eth0) can cause many issues, reverting the MTU to 1500 is likely what solved the problem.
BTW, the Fix Common Problems plugin is great for detecting configuration issues like this, it would have sent you here:
https://forums.unraid.net/topic/120220-fix-common-problems-more-information/page/2/#comment-1167702
It was blank so 1500 was the default. I just input 1500, applied, erased 1500, applied so it would go to auto.
-
1
-
-
blacktower-diagnostics-20231129
Diags attached and yes, the array is getting full. Cache pool upgrade and capacity upgrade is underway.
-
Ok, this might be solved.
Upgraded to 6.12.5 which seemed to help a tiny bit.
Discovered errors in my Network settings though.
1. Somehow port 2 on my 4 port Dell R720xd Ethernet was 'Up' but there's no cable...perhaps all the DHCP requests etc. it was probably doing?
2. ipv6 was enabled so I changed it to ipV4 only.
3. Rearranged DNS to 8.8.8.8 first and 1.1.1.1 second and 8.8.4.4 as third. Shuffled.
4. Put 1500 in as MYU, applied, erased and applied smaking sure it's Auto.
As of this moment, GUI is alo faster. Fingers crossed.
-
I'll attach diags when I can get them.
Think it's this process at over 100% CPU?
For instance, backing up USB drive takes forever. I wonder if the USB drive is failing.
Or, perhaps one of the disks is failing but there have been no SMART issues. Diskspeed seems to have trouble.
Real head scratcher. No errors just slow operation. Also at one point nginx wasn't running but nginx start started it.
Next time, I've got the USB to SATA key that gives me a GUID instead of a USB stick. I've had more problems with those...
Unraid, please just let us use a small SSD soon for boot. They're cheap.
-
Still having this issue.
One Clue? Web GUI said it was still unmounting but it shut down clean fine via GUI when I pressed that button in the GUI after about 10 mins.
Shut down so I could add an NVME.
root@BlackTower:~# losetup
NAME SIZELIMIT OFFSET AUTOCLEAR RO BACK-FILE DIO LOG-SEC
/dev/loop1 0 0 1 1 /boot/bzfirmware 0 512
/dev/loop0 0 0 1 1 /boot/bzmodules 0 512
root@BlackTower:~# umount /var/lib/docker
umount: /var/lib/docker: not mounted.
root@BlackTower:~#
-
Have had unlcean shutdowns in the past. Changed timeout to 300 seconds.
This time I also stopped the Dockers before a reboot.
Not sure if it's fixed but here's the diags.
blacktower-diagnostics-20231115-2027.zip
Getting error -200 on diag upload but I have the one before reboot.
-
From Intel Doc "Intel® Omni-Path Performance Tuning"
--quote--
Avoid acpi_pad Consuming CPU Resources
acpi_pad is a module ACPI Processor Aggregator Driver. It is supposed to handle high core count processor power management. Unfortunately, the driver can cause a system to run acpi_pad and consume 100% of each core.
A simple workaround is to blacklist acpi_pad by adding the following line to /etc/modprobe.d/blacklist.conf:
blacklist acpi_pad
This will keep the acpi_pad driver from loading at boot time.
If you need to remove acpi_pad without a reboot you can use:
modprobe –r acpi_pad
This removes the driver. However, note that the driver will restart upon a reboot.
--end quote--
modprobe –r acpi_pad worked.
modprobe acpi_pad if you want to reload it.
IMHO, this is an old Linux bug.
CPU E2697v2
-
On 10/15/2023 at 1:32 PM, Squid said:
Uninstall / reinstall
That's what it came down too. Install failed but apparently uninstalled then it asked me to install it.
-
-
Edit: Finally unounted and was able to stop the array. Restarted to 'fix' acpi_pad multiple processes consuming all of the CPU resources.
Waiting on other disk share unmounts but it did unmount the docker image finally.
The acpi_pad sucking up all CPUs is perhaps why THAT is taking so long.
Still mounted:
NAME SIZELIMIT OFFSET AUTOCLEAR RO BACK-FILE DIO LOG-SEC
/dev/loop1 0 0 1 1 /boot/bzfirmware 0 512
/dev/loop0 0 0 1 1 /boot/bzmodules 0 512
-
-
Can't get to GUI for diagnostics and not savvy enough for CMD process to do that but I'll try to find it later and update.
-
Similar situation.
SMB copy from Mac OS to Unraid share, flipped off cache mid copy and copy was not interrupted. HTH.
Was doing a copy via MC from an NVME to an /mnt share of 250+ GB with a 120GB cache, so it filled it. Aborted that, changed share to not use cache and re-did.
-
Doesn't OS X have SMB 3? Wouldn't connect with SMB 3 specified as min.
-
Specs:
Unraid 6.12.4
Cheap 120GB SSD Cache Disk
SMB User Shares
OS X 12.6.8
1Gb Ethernet
10GB Single File in about 2 mins.
Filled the RAM first of course, file is on the cache SATA SSD.
Used these from the Samba Wiki https://wiki.samba.org/index.php/Configure_Samba_to_Work_Better_with_Mac_OS_X
#Changes by Rob
server min protocol = SMB2
ea support = yes
#Fruit Settings
[Global]
vfs objects = fruit streams_xattr
fruit:metadata = stream
fruit:model = MacSamba
fruit:posix_rename = yes
fruit:veto_appledouble = no
fruit:nfs_aces = no
fruit:wipe_intentionally_left_blank_rfork = yes
fruit:delete_empty_adfiles = yes[TimeMachineBackup]
fruit:time machine = yes
#End changes by Rob -
2.10.6
Some drives don't appear to bench.
Stats plugin doesn't show activity, but drive is online, smart green etc. and hdparm works.
/dev/sdf:
Timing cached reads: 14826 MB in 1.99 seconds = 7443.88 MB/sec
Timing buffered disk reads: 500 MB in 3.01 seconds = 166.26 MB/sec
root@BlackTower:/mnt/disk5# sudo hdparm -Tt /dev/sdm/dev/sdm:
Timing cached reads: 18884 MB in 1.99 seconds = 9490.98 MB/sec
Timing buffered disk reads: 552 MB in 3.00 seconds = 183.74 MB/sec
But on those drives the Diskspeed webgui sits there.
WD40EFZX drives
Clicking Abort just also sits there.
/sdh works as expected. All drives are data drives in the array.
Nothing in logs. Over 30 mins old and last log entries:
WARNING: Use --illegal-access=warn to enable warnings of further illegal reflective access operations
WARNING: All illegal access operations will be denied in a future release
10-Jul-2023 18:00:42.061 INFO [main] org.apache.coyote.AbstractProtocol.start Starting ProtocolHandler ["http-nio-8888"]
10-Jul-2023 18:00:42.075 INFO [main] org.apache.catalina.startup.Catalina.start Server startup in [14279] millisecondsAlso rescanned controller, devs hadn't changed.
-
Similar.
06-07-2023 22:17Unraid Parity-CheckNotice [BLACKTOWER] - Parity-Check finished (0 errors)Canceledwarning
I didn't cancel it but my server locked (no hardware issues noted yet, but my usb stick was corrupted a couple of weeks ago)
Dirty bit set 'some corruption' warnings etc.
It appears to me, yet AGAIN that the USB stick is the weak link in Unraid.
For God's sake, please add some sort of redundancy at least to this weak link, say, it checks the USB stick against 2 other mirror files of it on the disk as a boot stick check.
[Support] binhex - MinecraftServer
in Docker Containers
Posted
I could not access the webgui with the error that there is no minecraft screen to attach to. I had forgotten the 'M' in the Dockert emplate settings for memory amount.
Server appeared to start but really didn't I presume.