PoMpIs Posted Thursday at 05:56 PM Share Posted Thursday at 05:56 PM (edited) Holaaa No soy muy asiduo por aqui, pero a ver si @JorgeB me lee por aqui, que es el que mas leo y entiende mucho de esto, que me cuesta expresarme en ingles ... Vamos al lio... Tengo un sistema con un ryzen 1600AF, placa MSI B450 Gaming Plus Max 32 gigas de ram y un porrón de discos para multimedia... Siempre funciono bien con UNRAID, pero cuando pase a 6.12.13, cambie los discos de cache a ZFS-Mirror, son 6 SSDs de distinto tamaño, pero en parejas.. es decir: -2x120Gb -2x250Gb -2x480Gb y quedaban tal que asi: Una vez cambiado el formato y vuelto a mover todo a la cache, todo parecia ir bien, pero en menos de 24 horas veo esto en el LOG: Y al final acababa por colgarse el ordenador, no respondia a nada.... Como ya tenia ganas de probar UNRAID 7 beta 2... y como tengo dos licencias PRO, pues cree otro pendrive con la otra key y meti la configuración que tenia en pendrive con UNRAID 6.12.13... Funciono perfectamente y todo parecia ir bien, pero al rato volvio a saltar esto: y si no reiniciaba en menos de 30 minutos se acababa colgando el server... Al final se me corrompio la imagen Docker.img, que estaba en formato BTRFS dentro del pool ZFS.... ya puestos, borre la imagen y cambie el Docker.img a XFS.... reinstalé todos los dockers y volvia a funcionar todo bien... pero al dia siguiente por la mañana ya volvia a tener esos errores en el LOG del sistema .... Ya le estaba echando la culpa a que el propio ordenador tuviera daño fisico de hardware, le probe varios kits de ram y seguia igual, baje la frecuencia de la CPU 1Ghz y subi el vCore y seguia fallando.... asi que el equipo funciona bien y sin fallos todo por defecto, asi que estaba claro que era algo del sistema operativo UNRAID... Al final, probe a cambiar el Docker.img a "Directorio" ... y ahi si que ya empezaron muchisimos mas problemas... con los discos de cache en ZFS, era imposible volver a reinstalar los docker, salian esos errores a destajo en el LOG, y no era capaz de instalar mas de 3 contenedores sin que se colgara el server... Al ver eso, me dije, a ver si va ser cosa de ZFS con estos discos o algo asi... Con lo cual, volvi a pasar mis discos de cache a BTRFS Raid1c3....... Y todos mis problemas se han acabado... no mas mensajes de error en el LOG no mas cuelgues nada de nada.. una maravilla y 100% estable.... Y sigo teniendo el Docker.img como Directorio en vez de imagen, y van super rápido los contenedores, tanto arrancandolos como instalandolos.. Me gustaria saber porque ZFS provoca esto en mi sistema???... porque me gustaba mucho usar ZFS_Master para las instantaneas... con BTRFS lo estoy haciendo a mano, porque el plugin que hay para hacerlas "snapshots" no se puede usar en UNRAID 7 Beta 2, porque deja el pool BTRFS en modo lectura y no queda mas remedio que vaciarlo y volver a formatearlo Muchas garcias Un Saludo Edited Thursday at 05:58 PM by PoMpIs Quote Link to comment
JorgeB Posted Thursday at 06:15 PM Share Posted Thursday at 06:15 PM The second call trace from v7 is harmless, but the symptoms look more like a RAM problem, start by posting the diagnostics to see the hardware used/config. 1 Quote Link to comment
PoMpIs Posted Thursday at 06:32 PM Author Share Posted Thursday at 06:32 PM (edited) 20 hours ago, JorgeB said: The second call trace from v7 is harmless, but the symptoms look more like a RAM problem, start by posting the diagnostics to see the hardware used/config. It can't be a RAM problem, I've tried 3 different new kits and it keeps showing up in the LOG until the server crashes.. I've run memtest, stability tests on the CPU and there's no hardware problem.... It's only when using cache disks in ZFS Once the CACHE disks are changed to BTRFS, none of that happens anymore, everything works 100% stable Thank you so much Edited Friday at 02:51 PM by PoMpIs Quote Link to comment
JorgeB Posted Friday at 07:48 AM Share Posted Friday at 07:48 AM Where you also using a docker folder with zfs? 1 Quote Link to comment
PoMpIs Posted Friday at 08:08 AM Author Share Posted Friday at 08:08 AM 2 minutes ago, JorgeB said: Where you also using a docker folder with zfs? With Docker.img-XFS and the cache disks in ZFS, the errors were the least likely to occur, but within 24 hours of operation, those errors would appear again in the LOG. With Docker.img-BTRFS and the cache disks in ZFS, I also saw those errors in the LOG in less than 24 hours With Docker-Folder, and the cache disks in ZFS, it gave me a lot of problems, I could no longer reinstall my dockers, those errors would appear in the LOG immediately. Once I moved the cache disks to BTRFS, I can use Docker.img in any format and those errors have not appeared in the LOG again, and the computer works without any problems. Now it is working wonderfully with Docker-Folder and the cache disks in BTRFS, those errors have not appeared in the LOG again, and everything works at 100%. Quote Link to comment
JorgeB Posted Friday at 09:47 AM Share Posted Friday at 09:47 AM 1 hour ago, PoMpIs said: With Docker-Folder, and the cache disks in ZFS, it gave me a lot of problems Theis is a known issue, see the release notes. It should not be a problem with an image, but the crash snippet you posted is not very helpful, you can enable the syslog server and retry, then post that after a new crash. 1 Quote Link to comment
PoMpIs Posted Friday at 10:24 AM Author Share Posted Friday at 10:24 AM (edited) 37 minutes ago, JorgeB said: Theis is a known issue, see the release notes. It should not be a problem with an image, but the crash snippet you posted is not very helpful, you can enable the syslog server and retry, then post that after a new crash. I imagined that with "Docker-Folder there must be some problem having the cache in ZFS, because I couldn't even reinstall 3 containers without the system failing... The problem is that in my system it ends up failing in the same way even if I use "Docker.img-XFS or Docker.img" and the cache disks in ZFS... it takes longer to fail, but it fails.. As I have two licenses, I will try with other hardware in case it is a problem of incompatibility of the motherboard with UNRAID and ZFS. This system with BTRFS is super stable, it has been used for more than a year as a 24/7 permaseed and records from 13 CCTV cameras with shinobi, and Jellyfin, sonar, radarr, immich, nextcloud, and many more, and it never failed, it started to fail as soon as I moved the cache disks to ZFS in 6.12.13... so for now I will leave it in BTRFS. Maybe Could it be that the disks I use for cache are on an ASMEDIA 1064 controller and cause problems with UNRAID when formatting the disks in ZFS? Thank you very much @JorgeB Edited Friday at 10:24 AM by PoMpIs Quote Link to comment
JorgeB Posted Friday at 11:15 AM Share Posted Friday at 11:15 AM 50 minutes ago, PoMpIs said: Maybe Could it be that the disks I use for cache are on an ASMEDIA 1064 controller and cause problems with UNRAID when formatting the disks in ZFS? Seems unlikely, many users using those controllers with zfs, including myself. 1 Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.