Upgraded one system already. No issues with the upgrade itself, but I'm currently running into two problems.
One of those problems has occurred before, starting with some build around 4000 (not entirely sure here, sorry)but was gone for some builds before the name change (please note that I don't update regularly, just when the NAS is active anyway).
The reoccurring problem is that during SMB transfers with lots of files the performance randomly drops after some time.
Sometimes the performance goes up again just to drop a bit later again.
What I could see is that the performance drops start to happen when the NAS is starting to use swap.
Before swap is used (right after startup, just when RAM is available) performance is high as expected.
I can see some log entries (see below) for swap and smbd but not sure if those apply here to the problem, since there are not always log entries when the performance drops.
Before I could get rid of the "swap" just by setting "vfs.zfs.arc_max" to "28G". Currently I'm at 23G but still experience swapping.
A new problem starting with this build are (again random) drops of the UPS connection.
The UPS itself is connected to my router. The router runs apcupsd (current version) and provides a server for all devices in the network.
The UPS is configured as "Master" with driver "apcupsd-ups" and the port is the IP address of my router. This worked all the time before.
Other clients connected to the UPS server do not report any connection loss, only the NAS. Right after a loss the connection is re-established.
System specs:
XigmaNAS x64 embedded
CPU: Intel(R) Xeon(R) CPU E3-1260L v5 @ 2.90GHz
Mainboard: Supermicro X11SSM-F BIOS 2.1a (yeah, not the current version, but every update is a major PITA and nothing changed between the updates before)
32GB RAM
8x 8TB Seagate HDD (on onboard-SATA controller) as a RAIDZ2
System installed on SanDisk Ultra Fit 16GB USB 3.0 Stick
1GB swap on said USB stick
Log entries with swap or smbd:
Code: Select all
Feb 5 07:51:41 MyNAS smbd[2223]: inotify_rm_watch returned Invalid argument
Feb 5 07:51:41 MyNAS smbd[2223]: [2019/02/05 07:51:41.768458, 1] ../source3/smbd/notify_inotify.c:369(watch_destructor)
Feb 5 07:51:41 MyNAS smbd[2223]: inotify_rm_watch returned Invalid argument
Feb 5 07:51:41 MyNAS smbd[2223]: [2019/02/05 07:51:41.721482, 1] ../source3/smbd/notify_inotify.c:369(watch_destructor)
Feb 5 07:51:41 MyNAS smbd[2223]: inotify_rm_watch returned Invalid argument
Feb 5 07:51:41 MyNAS smbd[2223]: [2019/02/05 07:51:41.559532, 1] ../source3/smbd/notify_inotify.c:369(watch_destructor)
Feb 5 07:46:47 MyNAS smbd[2223]: inotify_rm_watch returned Invalid argument
Feb 5 07:46:47 MyNAS smbd[2223]: [2019/02/05 07:46:47.163607, 1] ../source3/smbd/notify_inotify.c:369(watch_destructor)
Code: Select all
Feb 5 01:27:58 MyNAS kernel: swap_pager: indefinite wait buffer: bufobj: 0, blkno: 22357, size: 16384
Feb 5 01:27:58 MyNAS kernel: swap_pager: indefinite wait buffer: bufobj: 0, blkno: 1117, size: 8192
Feb 5 01:27:58 MyNAS kernel: swap_pager: indefinite wait buffer: bufobj: 0, blkno: 2098, size: 4096
Feb 5 01:27:30 MyNAS kernel: swap_pager: indefinite wait buffer: bufobj: 0, blkno: 2083, size: 4096
Log entries from the UPS outages (example):
Code: Select all
Feb 5 18:35:10 MyNAS upssched-cmd: Communications with the UPS apc1500@localhost are established
Feb 5 18:35:10 MyNAS upsmon[2060]: Communications with UPS apc1500@localhost established
Feb 5 18:35:05 MyNAS upssched-cmd: Communications with the UPS apc1500@localhost are lost
Feb 5 18:35:05 MyNAS upsmon[2060]: Communications with UPS apc1500@localhost lost
Feb 5 18:35:04 MyNAS upsmon[2060]: Poll UPS [apc1500@localhost] failed - Data stale
Feb 5 18:35:04 MyNAS upsd[2009]: UPS [apc1500] data is no longer stale
Feb 5 18:35:04 MyNAS upsd[2009]: Data for UPS [apc1500] is stale - check driver
Any more info I can provide to help finding the cause of the issue?
)