optiman

I changed my NTP server to us.pool.ntp.org no more errors, or at least not yet.

? Thank you!

battery is good and power has not been off. Server up time is over 30 days and this error only occurred once in the log. So I guess is happens about once a month. My time settings are default. Do you recommend different time settings?

I did not change anything from defaults. What do you recommend?

@limetech can you please see below errors and advise on how to resolve? Aug 17 17:35:32 Tower nginx: 2018/08/17 17:35:32 [error] 10387#10387: nchan: A message from the past has just been published. Unless the system time has been adjusted, this should never happen. Aug 17 17:35:32 Tower nginx: 2018/08/17 17:35:32 [error] 10387#10387: nchan: A message from the past has just been published. Unless the system time has been adjusted, this should never happen.

@juanamingo so at the end of the day, your rig was not running TRIM on the ssd's attached to your 9305 HBA. I too will just move the ssd to a mb sata port. I was hoping I could avoid using the mb and have everything on the 9305. I guess the only other option is to try the fw update, but I didn't see anything in the list of fixes that would help. Well, the other guy did also suggest to just NOT run the TRIM plugin, so that is a option too. I have it disabled right now.

@limetech can you please see my above posts on this page about testing TRIM and offer advice on how I should proceed? Also see my post here

@limetech Can you please see my first and second post and help me with these errors? @pwm Thank you

Chrome on my pc. No pop up, looks like plain text.

Cool, thank you.

The ssd drives that I want to run the TRIM on are the cache drives. I do not have any ssd drives that are used for data. Some users put the ssd drives on the motherboard sata controller to address TRIM issues. I would like to use my 9305 for all drives.

thank you Trurl! @Limetech - any help here with these errors? How can I resolve these so I have no errors?

I want to try to get this working but I could use some help from the smart people on this forum to test. Given juanamingo and me have the same LSI sata controller, both are using Samsung evo ssd drives connected to the 9305, and we are running same version of unraid. Yet his TRIM is working fine on his rig, and the last time I tried enabling this on my rig, it caused a bunch of errors, disk 1 redballed and many drives showed 1 error in the webgui Main page. Since I disabled this plugin, no errors at all. What is the best way to test TRIM without putting my data at risk? Should I start the array in Maintenance Mode, or have the array Stopped while I test?

I was hoping that LT would jump in here and help us figure out what these errors are, but no luck so far.

Thanks, very nice of you. I have to agree with you on fw update. Looking at what the fw update fixes, none have to do with trim or hard drive compatibility, so not likely the update would do much for me either. Never did get a response from LSI re my support ticket. I guess I just need to try to turn it on again and see what happens, but like you, I don't have time to mess with a glitchy system either. Thanks again!

ok, thanks for that information. I'm now at a loss as to why my setup does not like the trim command. We both have the same HBA and we both have samsung evo SSD's for cache. I'm afraid to try it again, given last time I lost a drive and several others showed an error on the main screen gui. I assume you have the latest version of unraid, 6.5.3? Perhaps it has something to do with my mb. I still have not heard back from LSI support. Really bad customer service via email. I guess I will have to call them.

I submitted a support ticket to Broadcom and after one week - no response. Not too impressed there. Again - anyone out there running this controller with a SSD drive attached and using this TRIM plugin with success? What is the down side of not running TRIM? I can leave the SSD drive connected to my 9305, but no TRIM. It has crashed my sever twice, so I have disabled this plugin.

The specs say that the SAS3224 does support SSD, but no mention of TRIM compatibility. I have opened a support ticket with Broadcom specifically asking if this controller supports SSD's and TRIM. I'll report back once I get the official answer. I would like to hear from anyone who has this card and has SSD's connected to it, and of course has the TRIM plugin working.

Can anyone confirm if the LSI 9305-24i HBA works with the SSD TRIM plugin? I have a Samsung 860 EVO 1tb as my cache and I'm having issues trying to run the TRIM command. I read some where that older LSI cards did not like the TRIM command. The fix back then was to use the motherboard sata ports, but I hope I don't have to do that.

Did you get a chance to upgrade your firmware on your 9305-24? Do you happen to have an SSD drive connected to it? If yes, any issues and does TRIM work or do you need to keep that disabled? I ask because I enabled TRIM and had some errors and then disk 1 red balled. Before I turn TRIM back on, I want to be sure our 9305's are compatible with TRIM. I read in another thread that some LSI HBA cards do not support TRIM with SSD drives. I also was thinking maybe the new firmware upgrade would fix it.

@ Squid - sent you a message ?

I installed a samsung 860 evo ssd drive a few weeks ago (working good) and then had a notification from the Fix Common problems that I need to enable Trim, so I did. I set it to run monthly on the 15th, so it tried to run last night. Today I see this error in my log: Jul 15 00:02:01 Tower kernel: sd 1:0:8:0: [sdj] tag#27 UNKNOWN(0x2003) Result: hostbyte=0x00 driverbyte=0x08 Jul 15 00:02:01 Tower kernel: sd 1:0:8:0: [sdj] tag#27 Sense Key : 0x5 [current] Jul 15 00:02:01 Tower kernel: sd 1:0:8:0: [sdj] tag#27 ASC=0x21 ASCQ=0x0 Jul 15 00:02:01 Tower kernel: sd 1:0:8:0: [sdj] tag#27 CDB: opcode=0x42 42 00 00 00 00 00 00 00 18 00 Jul 15 00:02:01 Tower kernel: print_req_error: critical target error, dev sdj, sector 1953277921 Jul 15 00:02:01 Tower root: /var/lib/docker: 34.9 GiB (37411815424 bytes) trimmed Given the word Critical is there, looks important. I have no SMART errors on any of my drives, webgui shows 0 errors for all drives and I ran Scrub and it had zero errors. Is this related to the latest release?

@ Limetech - any help here with these errors? How can I resolve these so I have no errors?

I found a couple more errors in the syslog. Jul 5 04:29:24 Tower kernel: diskload cpuset=/ mems_allowed=0 Jul 5 04:29:24 Tower kernel: CPU: 5 PID: 10492 Comm: diskload Not tainted 4.14.49-unRAID #1 Jul 5 04:29:24 Tower kernel: Hardware name: Supermicro Super Server/X11SSM-F, BIOS 2.0a 03/09/2017 Jul 5 04:29:24 Tower kernel: Call Trace: In the efforts to see if a clean syslog can be true, I post here again, and this time I include full syslog file. Thank you syslog.txt

Server is running great and no issues with performance, or anything else. However I did notice a few log entries that I don't recall seeing the past. They may have nothing to do with 6.5.3. Server has been up for about 2 days and 9 hours, and I see 3 entries in my syslog: Jun 29 03:41:13 Tower nginx: 2018/06/29 03:41:13 [error] 10391#10391: *296946 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: 127.0.0.1, server: , request: "POST /webGui/include/DeviceList.php HTTP/1.1", upstream: "fastcgi://unix:/var/run/php5-fpm.sock:", host: "localhost", referrer: "http://localhost/Main" Jun 29 03:41:13 Tower php-fpm[10365]: [WARNING] [pool www] child 21260 exited on signal 7 (SIGBUS) after 264.591907 seconds from start Jun 29 20:40:49 Tower nginx: 2018/06/29 20:40:49 [error] 10391#10391: *457950 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: 127.0.0.1, server: , request: "POST /webGui/include/DeviceList.php HTTP/1.1", upstream: "fastcgi://unix:/var/run/php5-fpm.sock:", host: "localhost", referrer: "http://localhost/Main" Jun 29 20:40:49 Tower php-fpm[10365]: [WARNING] [pool www] child 9358 exited on signal 7 (SIGBUS) after 337.768022 seconds from start Jun 30 16:16:14 Tower nginx: 2018/06/30 16:16:14 [error] 10391#10391: *618350 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: 127.0.0.1, server: , request: "POST /webGui/include/DeviceList.php HTTP/1.1", upstream: "fastcgi://unix:/var/run/php5-fpm.sock:", host: "localhost", referrer: "http://localhost/Main" Jun 30 16:16:14 Tower php-fpm[10365]: [WARNING] [pool www] child 17068 exited on signal 7 (SIGBUS) after 289.569545 seconds from start I also see the following warnings in my log: Jun 28 07:46:17 Tower kernel: ACPI: Early table checksum verification disabled Jun 28 07:46:17 Tower kernel: spurious 8259A interrupt: IRQ7. Jun 28 07:46:17 Tower kernel: random: 7 urandom warning(s) missed due to ratelimiting Jun 28 07:46:18 Tower rpc.statd[1769]: Failed to read /var/lib/nfs/state: Success Jun 28 07:46:18 Tower rpc.statd[1769]: Initializing NSM state Jun 28 07:46:18 Tower sshd[1787]: Server listening on 0.0.0.0 port 22. Jun 28 07:46:18 Tower sshd[1787]: Server listening on :: port 22. Jun 28 07:46:39 Tower sshd[1787]: Received signal 15; terminating. Jun 28 07:46:39 Tower acpid: client connected from 10225[0:0] Jun 28 07:46:39 Tower acpid: 1 client rule loaded Jun 28 07:46:40 Tower sshd[10242]: Server listening on 0.0.0.0 port 22. Jun 28 07:46:40 Tower sshd[10242]: Server listening on :: port 22. Jun 28 07:46:41 Tower avahi-daemon[10335]: WARNING: No NSS support for mDNS detected, consider installing nss-mdns! Other than these few warnings / errors, my log is clean. Do any of these look serious or important? Thanks!