[Help] RAID Array rebuilds on each reboot

[Misto]

Hello,

I recently lost power and my Terramaster f4-423 didn't shut down cleanly. When I rebooted the NAS, it took around 8 hours to boot into a useable state (being able to SSH or sign into the web interface). All the data was still available, although response times were slow. I shut it down again (cleanly) and the same thing happened on the next reboot. I did some digging this time and found a few issues when running dmesg and inspecting the /var/log/messages file.

From dmesg:

Code: Select all

[   40.963879] [1559]: scst: scst_unregister_device:1163:***ERROR***: SCST device for SCSI device 4:0:0:0 not found
[   53.503450] md: md8 stopped.
[   53.543895] md/raid1:md8: active with 1 out of 72 mirrors
[   53.573626] md8: detected capacity change from 0 to 3997696
[   53.794545] md: delaying recovery of md8 until md9 has finished (they share one or more physical units)
[   54.240729] Adding 1998844k swap on /dev/md8.  Priority:-2 extents:1 across:1998844k 
[   54.353185] md: md2 stopped.
[   54.397425] md/raid:md2: not clean -- starting background reconstruction
[   54.397444] md/raid:md2: device sdc4 operational as raid disk 0
[   54.397447] md/raid:md2: device sda4 operational as raid disk 3
[   54.397449] md/raid:md2: device sdb4 operational as raid disk 2
[   54.397451] md/raid:md2: device sdd4 operational as raid disk 1
[   54.397681] md/raid:md2: raid level 5 active with 4 out of 4 devices, algorithm 2
[   54.397701] md2: detected capacity change from 0 to 93692725968
[   54.397740] md: delaying resync of md2 until md8 has finished (they share one or more physical units)
[   54.509428] md: md0 stopped.
[   54.510526] md/raid1:md0: active with 2 out of 2 mirrors
[   54.510540] md0: detected capacity change from 0 to 3904634880
[   54.510555] md: delaying recovery of md8 until md2 has finished (they share one or more physical units)
[   54.510557] md: delaying resync of md2 until md9 has finished (they share one or more physical units)
[   65.912015] md: md9: recovery done.
[   65.923387] md: resync of RAID array md2
[  133.166172] flashcache: flashcache-3.1.1 initialized
[  153.042711] EXT4-fs (dm-0): mounted filesystem with ordered data mode. Opts: data=ordered,user_xattr,acl,usrjquota=quota.user,grpjquota=quota.group,jqfmt=vfsv1. Quota mode: journalled.
[20264.793529] quotacheck (2679) used greatest stack depth: 12800 bytes left
[20340.764443] udevd[3978]: starting version 3.2.4

From /var/log/messages:

Code: Select all

2026-03-20 18:45:51|daemon|127.0.0.1|TNAS|-|err|nmbd| [2026/03/20 18:45:51.247931,  0] ../../source3/nmbd/nmbd.c:901(main)
2026-03-20 18:45:51|daemon|127.0.0.1|TNAS|-|err|nmbd|   nmbd version 4.15.6 started.
2026-03-20 18:45:51|daemon|127.0.0.1|TNAS|-|err|nmbd|   Copyright Andrew Tridgell and the Samba Team 1992-2021
2026-03-20 18:45:51|kern|127.0.0.1|TNAS|-|info|kernel|[20361.988977] sed (7237) used greatest stack depth: 12664 bytes left
2026-03-20 18:45:51|syslog|127.0.0.1|TNAS|-|info|rsyslogd| [origin software="rsyslogd" swVersion="8.2010.0" x-pid="7224" x-info="https://www.rsyslog.com"] start
2026-03-20 18:45:56|syslog|127.0.0.1|TNAS|-|info|rsyslogd| [origin software="rsyslogd" swVersion="8.2010.0" x-pid="7303" x-info="https://www.rsyslog.com"] start

2026-03-20 21:56:23|user|127.0.0.1|TNAS|-|notice|MDADM-MONITOR| [Rebuild40]: raid /dev/md2 device is Rebuild40
Mar 20 21:56:23 TNAS mdadm[4053]: Rebuild40 event detected on md device /dev/md2
Mar 20 22:06:49 TNAS -- MARK --
Mar 20 22:17:52 TNAS smbd[12211]: [2026/03/20 22:17:52.614199,  0] ../../source3/passdb/pdb_smbpasswd.c:1252(build_sam_account) 
Mar 20 22:17:52 TNAS smbd[12211]:   build_sam_account: smbpasswd database is corrupt!  username admin with uid 0 is not in unix passwd database! 
2026-03-20 22:22:15|auth|127.0.0.1|TNAS|-|info|sshd| Accepted keyboard-interactive/pam for from  ssh2
Mar 20 22:37:26 TNAS smbd_audit: [2026/03/20 22:37:26.803774,  0] ../../source3/passdb/pdb_smbpasswd.c:1252(build_sam_account) 
Mar 20 22:37:26 TNAS smbd_audit:   build_sam_account: smbpasswd database is corrupt!  username admin with uid 0 is not in unix passwd database! 

Again, once it booted up, all the data was there (slow but correct) and all the drives detect as healthy. Is there anything I can do to have it recognize as healthy on boot and not have to rebuild each time?
Some info about the RAID:
Using TRAID with 4 HDDs - 24tb, 18tb, 2x 16tb

Let me know if there is any other info I could provide. Thanks for any help!

[CursaYang]

The hard drive data inconsistency may be caused by a previous power outage. Please generate and download the log report in TOS > Technical Support > Issue Report, then send it to our technical support email ([email protected]). We will analyze it.