March 11, 20215 yr Syslog attached. Detail of build is in my sig. Motherboard BMC chip died and has been replaced. Other than that and the upgrade to 6.9. Nothing out of the norm. Also has UPS if that makes any difference. thelibrary-syslog-20210311-0928.zip
March 11, 20215 yr Community Expert Assuming loop2 is the docker image, if you're not sure post the complete diagnostics, delete and recreate.
March 11, 20215 yr Author Docker deleted (had to reboot, that kills all the logs and what not right?) and rebuilt. I guess now we wait.
March 13, 20215 yr Author Ok here we go, It happened again. I reformatted the Cache disks after my post here. thelibrary-diagnostics-20210313-0116.zip
March 13, 20215 yr Community Expert Ideally 24H, but if there's a considerable problem it should be found quickly.
March 13, 20215 yr Author Anything wrong with taking two sticks at a time to another machine to test?
March 13, 20215 yr Community Expert You can do that, but if you don't find the problem should then still run it in the server as is.
March 13, 20215 yr Community Expert 10 minutes ago, DuzAwe said: Anything wrong with taking two sticks at a time to another machine to test? That will check that the RAM sticks themselves are not faulty. However, sometimes ram needs to be tested "in situ" in case there is any sort of bus loading issue affecting RAM. Also different motherboard/CPU combinations may have different maximum RAM clock rates they successfully support regardless of what the RAM itself has specified as its maximum clock rate.
March 13, 20215 yr Author Have been thinking. Since all this started I haven't been able to get the smart tests to run on the cache drives. Is that more indicative of the drives failing then the RAM being an issue? Memtest is two passes down with no errors.
March 14, 20215 yr Community Expert 18 hours ago, DuzAwe said: Since all this started I haven't been able to get the smart tests to run on the cache drives. You can't run SMART tests on NVMe devices. Let it run for 24H but if it didn't find any issues so far it likely won't, also note that memtest sometimes can't find issues even when there are some, but the problem can come from somewhere else.
March 14, 20215 yr Author 3.5 hours left on the test and its all clear. What should I gear up for next? Edited March 14, 20215 yr by DuzAwe
March 14, 20215 yr Community Expert Remove two DIMMs and reacreate the docker image, if it corrupts again try with just the other two, if still issues unlikely to be RAM related.
March 14, 20215 yr JFYI, I started dealing with this same issue on my own last night. Also after upgrading to 6.9.1 a day earlier.
March 14, 20215 yr Author Its worth noting in that case that this issue started with 6.9-RC2. My set up has not changed since it was built in Jan 2020 (better times) Its run stable up to the xmas period. With a few lock ups I'm starting to believe is related to logs filling the ram up. But Im not sure. Anyway I have caught the device loop2 issue and hope to help get to the bottom of it. Edited March 14, 20215 yr by DuzAwe
March 14, 20215 yr Author So new vdisk created and its immediately corrupt thelibrary-diagnostics-20210314-1413.zip
March 14, 20215 yr Author Got a lovely 503 for web gui on my reboot thelibrary-diagnostics-20210314-1427.zip
March 14, 20215 yr Author Not to jinx it but..........the next set of ram dimms seem to have everything stable...........logs attached just because. I have ordered ECC replacement sticks to hopefully rule this issue out permanently. Hopefully this is the end of this issue. Thanks for all the help. Ill report back should anything happen once I start using the system properly again. thelibrary-diagnostics-20210314-1533.zip
Archived
This topic is now archived and is closed to further replies.