Hi,
I am trying to debug a strange issue and I am at my wits' (and google skills) end.
Running most recent PVE on Debian Stretch (as per the Wiki) with full disk encryption and remote cryptroot unlock.
My host has four physical 8TB HDDs (sda - sdd); unmounted and used for direct passthrough and one SSD (sde) as root and VM storage, setup as per the original PVE installer as VG with 3 separate LVs for swap, root and data. The data volume is setup as a thinpool.
With the server at complete idle (no VMs or any tasks running) the kernel is accessing the disks exactly every 10 seconds, which prevents spindown and is also quite annoying when sitting next to the server
Not only the root device is affected, but also the four unmounted HDDs (all LUKS devices).
It also makes no difference if the devices are actively being passed through to a running VM or not.
Disabling pve-ha-lrm and pve-ha-crm via systemctl as per this thread seemingly only reduced the access intervals to the SSD, but not the HDDs.
A block dump shows lvs and vgs reading the same blocks on all devices every 10 seconds:
If I kill pmxcfs the strange access pattern stops immediately.
Any advice on what's causing this strange behavior and how to further debug and/or solve it?
Thanks and regards,
Heiner
I am trying to debug a strange issue and I am at my wits' (and google skills) end.
Running most recent PVE on Debian Stretch (as per the Wiki) with full disk encryption and remote cryptroot unlock.
Code:
uname -a
Linux pve 4.13.4-1-pve #1 SMP PVE 4.13.4-26 (Mon, 6 Nov 2017 11:23:55 +0100) x86_64 GNU/Linux
My host has four physical 8TB HDDs (sda - sdd); unmounted and used for direct passthrough and one SSD (sde) as root and VM storage, setup as per the original PVE installer as VG with 3 separate LVs for swap, root and data. The data volume is setup as a thinpool.
Code:
lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 1 7.3T 0 disk
sdb 8:16 1 7.3T 0 disk
sdc 8:32 1 7.3T 0 disk
sdd 8:48 1 7.3T 0 disk
sde 8:64 1 232.9G 0 disk
+-sde1 8:65 1 243M 0 part /boot
+-sde2 8:66 1 1K 0 part
+-sde5 8:69 1 232.7G 0 part
+-sda5_crypt 253:0 0 232.7G 0 crypt
+-b0x--pve--vg-root 253:1 0 18.6G 0 lvm /
+-b0x--pve--vg-swap 253:2 0 9.1G 0 lvm [SWAP]
+-b0x--pve--vg-data_tmeta 253:3 0 104M 0 lvm
¦ +-b0x--pve--vg-data-tpool 253:5 0 204.7G 0 lvm
¦ +-b0x--pve--vg-data 253:6 0 204.7G 0 lvm
¦ +-b0x--pve--vg-vm--100--disk--1 253:7 0 10G 0 lvm
¦ +-b0x--pve--vg-vm--101--disk--1 253:8 0 30G 0 lvm
¦ +-b0x--pve--vg-vm--101--disk--2 253:9 0 105G 0 lvm
+-b0x--pve--vg-data_tdata 253:4 0 204.7G 0 lvm
+-b0x--pve--vg-data-tpool 253:5 0 204.7G 0 lvm
+-b0x--pve--vg-data 253:6 0 204.7G 0 lvm
+-b0x--pve--vg-vm--100--disk--1 253:7 0 10G 0 lvm
+-b0x--pve--vg-vm--101--disk--1 253:8 0 30G 0 lvm
+-b0x--pve--vg-vm--101--disk--2 253:9 0 105G 0 lvm
With the server at complete idle (no VMs or any tasks running) the kernel is accessing the disks exactly every 10 seconds, which prevents spindown and is also quite annoying when sitting next to the server
Not only the root device is affected, but also the four unmounted HDDs (all LUKS devices).
It also makes no difference if the devices are actively being passed through to a running VM or not.
Code:
iotop -o -b -d 10
Total DISK READ : 0.00 B/s | Total DISK WRITE : 0.00 B/s
Actual DISK READ: 26.14 K/s | Actual DISK WRITE: 1225.97 B/s
TID PRIO USER DISK READ DISK WRITE SWAPIN IO COMMAND
5051 be/0 root 0.00 B/s 0.00 B/s 0.00 % 99.99 % [kworker/u17:1]
16 be/4 root 0.00 B/s 0.00 B/s 0.00 % 99.99 % [ksoftirqd/1]
28 be/4 root 0.00 B/s 0.00 B/s 0.00 % 99.99 % [ksoftirqd/3]
7 be/4 root 0.00 B/s 0.00 B/s 0.00 % 99.99 % [ksoftirqd/0]
22 be/4 root 0.00 B/s 0.00 B/s 0.00 % 99.99 % [ksoftirqd/2]
34 be/4 root 0.00 B/s 0.00 B/s 0.00 % 99.99 % [ksoftirqd/4]
40 be/4 root 0.00 B/s 0.00 B/s 0.00 % 99.99 % [ksoftirqd/5]
Disabling pve-ha-lrm and pve-ha-crm via systemctl as per this thread seemingly only reduced the access intervals to the SSD, but not the HDDs.
A block dump shows lvs and vgs reading the same blocks on all devices every 10 seconds:
Code:
echo 1 > /proc/sys/vm/block_dump
tail -f /var/log/kern.log | grep sd
Nov 25 10:23:40 pve kernel: [150901.104951] vgs(13121): READ block 0 on sda (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.105238] vgs(13121): READ block 15628052992 on sda (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.105358] vgs(13121): READ block 15628053152 on sda (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.105480] vgs(13121): READ block 0 on sda (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.105624] vgs(13121): READ block 8 on sda (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.105777] vgs(13121): READ block 0 on sda (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.111537] vgs(13121): READ block 0 on sdb (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.111796] vgs(13121): READ block 15628052992 on sdb (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.111915] vgs(13121): READ block 15628053152 on sdb (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.112024] vgs(13121): READ block 0 on sdb (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.112154] vgs(13121): READ block 8 on sdb (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.112334] vgs(13121): READ block 0 on sdb (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.112522] vgs(13121): READ block 0 on sdc (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.112753] vgs(13121): READ block 15628052992 on sdc (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.112885] vgs(13121): READ block 15628053152 on sdc (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.113034] vgs(13121): READ block 0 on sdc (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.113162] vgs(13121): READ block 8 on sdc (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.113337] vgs(13121): READ block 0 on sdc (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.113537] vgs(13121): READ block 0 on sdd (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.113751] vgs(13121): READ block 15628052992 on sdd (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.113865] vgs(13121): READ block 15628053152 on sdd (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.113966] vgs(13121): READ block 0 on sdd (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.114094] vgs(13121): READ block 8 on sdd (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.114260] vgs(13121): READ block 0 on sdd (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.114447] vgs(13121): READ block 0 on sde (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.114614] vgs(13121): READ block 497536 on sde1 (2 sectors)
Nov 25 10:23:40 pve kernel: [150901.114650] vgs(13121): READ block 497648 on sde1 (2 sectors)
Nov 25 10:23:40 pve kernel: [150901.114678] vgs(13121): READ block 0 on sde1 (2 sectors)
Nov 25 10:23:40 pve kernel: [150901.114795] vgs(13121): READ block 8 on sde1 (2 sectors)
Nov 25 10:23:40 pve kernel: [150901.114930] vgs(13121): READ block 0 on sde1 (4 sectors)
Nov 25 10:23:40 pve kernel: [150901.115141] vgs(13121): READ block 487894912 on sde5 (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.115266] vgs(13121): READ block 487895024 on sde5 (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.115374] vgs(13121): READ block 0 on sde5 (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.115479] vgs(13121): READ block 8 on sde5 (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.115632] vgs(13121): READ block 0 on sde5 (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.128351] lvs(13122): READ block 0 on sda (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.128543] lvs(13122): READ block 15628052992 on sda (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.128641] lvs(13122): READ block 15628053152 on sda (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.128739] lvs(13122): READ block 0 on sda (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.128835] lvs(13122): READ block 8 on sda (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.129001] lvs(13122): READ block 0 on sda (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.133275] lvs(13122): READ block 0 on sdb (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.133445] lvs(13122): READ block 15628052992 on sdb (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.133534] lvs(13122): READ block 15628053152 on sdb (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.133628] lvs(13122): READ block 0 on sdb (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.133723] lvs(13122): READ block 8 on sdb (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.133858] lvs(13122): READ block 0 on sdb (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.133988] lvs(13122): READ block 0 on sdc (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.134113] lvs(13122): READ block 15628052992 on sdc (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.134204] lvs(13122): READ block 15628053152 on sdc (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.134306] lvs(13122): READ block 0 on sdc (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.134401] lvs(13122): READ block 8 on sdc (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.134535] lvs(13122): READ block 0 on sdc (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.134675] lvs(13122): READ block 0 on sdd (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.134807] lvs(13122): READ block 15628052992 on sdd (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.134885] lvs(13122): READ block 15628053152 on sdd (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.134985] lvs(13122): READ block 0 on sdd (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.135087] lvs(13122): READ block 8 on sdd (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.135239] lvs(13122): READ block 0 on sdd (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.135374] lvs(13122): READ block 0 on sde (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.135571] lvs(13122): READ block 497536 on sde1 (2 sectors)
Nov 25 10:23:40 pve kernel: [150901.135604] lvs(13122): READ block 497648 on sde1 (2 sectors)
Nov 25 10:23:40 pve kernel: [150901.135631] lvs(13122): READ block 0 on sde1 (2 sectors)
Nov 25 10:23:40 pve kernel: [150901.135738] lvs(13122): READ block 8 on sde1 (2 sectors)
Nov 25 10:23:40 pve kernel: [150901.135867] lvs(13122): READ block 0 on sde1 (4 sectors)
Nov 25 10:23:40 pve kernel: [150901.136091] lvs(13122): READ block 487894912 on sde5 (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.136209] lvs(13122): READ block 487895024 on sde5 (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.136309] lvs(13122): READ block 0 on sde5 (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.136406] lvs(13122): READ block 8 on sde5 (8 sectors)
Nov 25 10:23:40 pve kernel: [150901.136581] lvs(13122): READ block 0 on sde5 (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.168980] vgs(13130): READ block 0 on sda (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.169232] vgs(13130): READ block 15628052992 on sda (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.169353] vgs(13130): READ block 15628053152 on sda (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.169469] vgs(13130): READ block 0 on sda (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.169597] vgs(13130): READ block 8 on sda (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.169761] vgs(13130): READ block 0 on sda (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.175689] vgs(13130): READ block 0 on sdb (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.175934] vgs(13130): READ block 15628052992 on sdb (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.176060] vgs(13130): READ block 15628053152 on sdb (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.176173] vgs(13130): READ block 0 on sdb (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.176305] vgs(13130): READ block 8 on sdb (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.176466] vgs(13130): READ block 0 on sdb (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.176628] vgs(13130): READ block 0 on sdc (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.176861] vgs(13130): READ block 15628052992 on sdc (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.176989] vgs(13130): READ block 15628053152 on sdc (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.177099] vgs(13130): READ block 0 on sdc (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.177228] vgs(13130): READ block 8 on sdc (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.177415] vgs(13130): READ block 0 on sdc (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.177570] vgs(13130): READ block 0 on sdd (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.177797] vgs(13130): READ block 15628052992 on sdd (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.177900] vgs(13130): READ block 15628053152 on sdd (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.178010] vgs(13130): READ block 0 on sdd (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.178134] vgs(13130): READ block 8 on sdd (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.178268] vgs(13130): READ block 0 on sdd (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.178430] vgs(13130): READ block 0 on sde (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.178607] vgs(13130): READ block 497536 on sde1 (2 sectors)
Nov 25 10:23:50 pve kernel: [150911.178639] vgs(13130): READ block 497648 on sde1 (2 sectors)
Nov 25 10:23:50 pve kernel: [150911.178669] vgs(13130): READ block 0 on sde1 (2 sectors)
Nov 25 10:23:50 pve kernel: [150911.178779] vgs(13130): READ block 8 on sde1 (2 sectors)
Nov 25 10:23:50 pve kernel: [150911.178928] vgs(13130): READ block 0 on sde1 (4 sectors)
Nov 25 10:23:50 pve kernel: [150911.179120] vgs(13130): READ block 487894912 on sde5 (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.179240] vgs(13130): READ block 487895024 on sde5 (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.179344] vgs(13130): READ block 0 on sde5 (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.179444] vgs(13130): READ block 8 on sde5 (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.179616] vgs(13130): READ block 0 on sde5 (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.192421] lvs(13131): READ block 0 on sda (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.192585] lvs(13131): READ block 15628052992 on sda (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.192667] lvs(13131): READ block 15628053152 on sda (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.192766] lvs(13131): READ block 0 on sda (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.192861] lvs(13131): READ block 8 on sda (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.193005] lvs(13131): READ block 0 on sda (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.197204] lvs(13131): READ block 0 on sdb (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.197362] lvs(13131): READ block 15628052992 on sdb (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.197443] lvs(13131): READ block 15628053152 on sdb (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.197583] lvs(13131): READ block 0 on sdb (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.197679] lvs(13131): READ block 8 on sdb (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.197859] lvs(13131): READ block 0 on sdb (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.198014] lvs(13131): READ block 0 on sdc (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.198150] lvs(13131): READ block 15628052992 on sdc (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.198243] lvs(13131): READ block 15628053152 on sdc (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.198341] lvs(13131): READ block 0 on sdc (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.198438] lvs(13131): READ block 8 on sdc (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.198549] lvs(13131): READ block 0 on sdc (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.198684] lvs(13131): READ block 0 on sdd (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.198858] lvs(13131): READ block 15628052992 on sdd (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.198956] lvs(13131): READ block 15628053152 on sdd (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.199078] lvs(13131): READ block 0 on sdd (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.199187] lvs(13131): READ block 8 on sdd (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.199323] lvs(13131): READ block 0 on sdd (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.199443] lvs(13131): READ block 0 on sde (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.199622] lvs(13131): READ block 497536 on sde1 (2 sectors)
Nov 25 10:23:50 pve kernel: [150911.199651] lvs(13131): READ block 497648 on sde1 (2 sectors)
Nov 25 10:23:50 pve kernel: [150911.199678] lvs(13131): READ block 0 on sde1 (2 sectors)
Nov 25 10:23:50 pve kernel: [150911.199786] lvs(13131): READ block 8 on sde1 (2 sectors)
Nov 25 10:23:50 pve kernel: [150911.199923] lvs(13131): READ block 0 on sde1 (4 sectors)
Nov 25 10:23:50 pve kernel: [150911.200135] lvs(13131): READ block 487894912 on sde5 (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.200256] lvs(13131): READ block 487895024 on sde5 (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.200355] lvs(13131): READ block 0 on sde5 (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.200454] lvs(13131): READ block 8 on sde5 (8 sectors)
Nov 25 10:23:50 pve kernel: [150911.200619] lvs(13131): READ block 0 on sde5 (8 sectors)
If I kill pmxcfs the strange access pattern stops immediately.
Any advice on what's causing this strange behavior and how to further debug and/or solve it?
Thanks and regards,
Heiner