Frühere Suchanfragen

Keine früheren Suchanfragen

Suchoptionen

Nur verfügbar, wenn angemeldet.
mastodontech.de ist einer von vielen unabhängigen Mastodon-Servern, mit dem du dich im Fediverse beteiligen kannst.
Offen für alle (über 16) und bereitgestellt von Markus'Blog

Verwaltet von:

Serverstatistik:

1,5 Tsd.
aktive Profile

mastodontech.de: ÜberStatusProfilverzeichnisDatenschutzerklärung

Mastodon: ÜberApp herunterladenTastenkombinationenQuellcode anzeigenv4.4.0-beta.2

#zfs

12 Beiträge11 Beteiligte1 Beitrag heute
Practical ZFS - Discussions
A drive in my zpool had a "transient error". Do you agree with chatgpt's assessment?

Hi,
About a month ago I noticed read errors on one drive in my zfs pool. The pool consists of 4 vdevs, each vdev is a 2 drive mirror.

zpool status reported that the previous scrub (early in May) had corrected some errors (I believe it was a few megabytes worth).
So I ran a extended SMART tests on the drives in the mirror, and one had some concerning results and a read failure. The other drive was fine, so I didn’t panic.
But, since I don’t have much experience with interpreting SMART errors, I asked chatgpt for help.

Here’s the chat (with more details about the SMART errors):

ChatGPT

ChatGPT - SMART UNC error explanation

Shared via ChatGPT

As you can see near the bottom of the chat, I have since moved all the drives to a new computer/server, in a new case with better cooling, and I also replaced all SATA cables in the process. After doing all this I ran a new extended SMART test, and the drive now reports no errors! (I realize that making all these changes together makes it harder to pinpoint what the exact cause of the errors was.)

I have also ordered two new drives, which at first I intended to use to replace the failing drive + its mirror buddy.
But, since the “failing” drive now reports that it’s fine, I’m considering just adding the new drives to the pool as a new vdev, to gain a whole lot more free space (which I need).
ChatGPT thinks the old drives should be fine, and its assessment seems solid (to me), but I also don’t want to blindly trust AI… and I’ve never experienced this kind of “transient error” before.

Any input / thoughts on this situation? Do you agree with ChatGPT’s assessment? Should I replace the drives, or keep them a while longer? Can I trust the successful SMART test? The two drives were bought together and added to the pool at the same time, and have been running more or less 24/7 for almost 3 years. (And yes, I’m gonna set up automated regular SMART tests + alerts asap, which is something I’ve neglected for far too long…)

Thanks in advance for any advice or input

2 posts - 2 participants

Read full topic

Practical ZFS · A drive in my zpool had a "transient error". Do you agree with chatgpt's assessment?Hi, About a month ago I noticed read errors on one drive in my zfs pool. The pool consists of 4 vdevs, each vdev is a 2 drive mirror. zpool status reported that the previous scrub (early in May) had corrected some errors (I believe it was a few megabytes worth). So I ran a extended SMART tests on the drives in the mirror, and one had some concerning results and a read failure. The other drive was fine, so I didn’t panic. But, since I don’t have much experience with interpreting SMART errors,...
#zfs
Practical ZFS - Discussions
Mountpoints and replicating multiple machines to a single zpool

I have a long-running Debian server (myserver)that’s had everything but the OS moved into a zpool (output of command redacted, file sizes won’t add up):

user1@myserver:~$ zfs list
NAME                     USED  AVAIL     REFER  MOUNTPOINT
tank                    3.02T   505G      104K  /tank
tank/data               1.50T   505G     6.28M  /tank/data
tank/data/home           511G   505G       96K  /tank/data/home
tank/data/home/user1     259G   505G      259G  /home/user1
tank/data/home/user2    30.8M   505G     10.7M  /home/user2
tank/data/vm             321G   505G      200G  /tank/data/vm
tank/ephemeral          51.6G   505G     4.40G  /tank/ephemeral
tank/reserved            700G  1.18T       96K  /tank/reserved

I’m replicating the server to another one but I’m also doing periodic backups to a USB HDD wd202501 which has a single zpool called wd202501 and a dataset called bkmyserver. I formatted the drive and backed up my server. No problems.

I also have a laptop (mylaptop) onto which I recently installed ZFSBootMenu and Ubuntu:

user1@mylaptop:~$ zfs list
NAME                      USED  AVAIL  REFER  MOUNTPOINT
zroot                     375G  1.39T   192K  none
zroot/ROOT               10.8G  1.39T   192K  none
zroot/ROOT/ubuntu        10.8G  1.39T  10.8G  /
zroot/home                364G  1.39T  7.62M  /home
zroot/home/user1          364G  1.39T   344G  /home/user1
zroot/home/user2          952K  1.39T   952K  /home/user2

I would like to also back up my laptop to wd202501, in a dataset called bkmylaptop.

I attached he drive to my laptop and ran zpool import:

user1@mylaptop:~$ sudo zpool import -d /dev/disk/by-id/ata-WDC_REDACTED-REDACTEDREDACTEDREDACTED-part1 wd202501 -N

Before I ran zfs mount -l wd202501/bkwilson I ran zfs list and saw this:

user1@mylaptop:~$ zfs list
NAME                                  USED  AVAIL  REFER  MOUNTPOINT
wd202501                             1.30T  2.21T   104K  /mnt/wd202501
wd202501/bkmyserver                  1.20T  2.21T   200K  /mnt/wd202501/bkmyserver
wd202501/bkmyserver/data             1.20T  2.21T  1.36M  /mnt/wd202501/bkmyserver/data
wd202501/bkmyserver/data/home         496G  2.21T    96K  /tank/data/home
wd202501/bkmyserver/data/home/user1   248G  2.21T   248G  /home/user1
wd202501/bkmyserver/data/home/user2  2.77M  2.21T  2.77M  /home/user2
wd202501/bkmyserver/data/vm           111G  2.21T  67.8G  /mnt/wd202501/bkmyserver/data/vm
wd202501/reserved                     100G  2.31T    96K  /mnt/wd202501/reserved
zroot                                 375G  1.39T   192K  none
zroot/ROOT                           10.8G  1.39T   192K  none
zroot/ROOT/ubuntu                    10.8G  1.39T  10.8G  /
zroot/home                            364G  1.39T  7.62M  /home
zroot/home/user1                      364G  1.39T   344G  /home/user1
zroot/home/user2                      952K  1.39T   952K  /home/user2

Both zroot/home/user1 and wd202504/bkmyserver/data/home/user1 have a mountpoint set to /home/user1. It does not appear that wd202504/bkmyserver/data/home/user1 overwrote zroot/home/user1, but I had not run zfs mount yet.

[The mountpoint for wd202501/bkmyserver/data/home points to /tank/data/home rather than /home but I’m not sure if that’s important for this question.]

My questions:

If I run zfs mount will wd202501/bkmyserver/data/home/user1 clobber zroot/home/user1 on mylaptop?

If I replicate mylaptop’s zroot to wd202501/bkmylaptop, will the two mountpoints conflict with each other? Will they conflict if I replicate mylaptop to my replication server?

Should I just not be doing these mountpoints in the first place?

Where do the mountpoint settings reside? Are they part of the dataset or the zpool?

Is it possible to keep the mountpoints on the source zpool but unset them on the backup zpool? Will they get re-set whenever I replicate a snapshot?

1 post - 1 participant

Read full topic

Practical ZFS · Mountpoints and replicating multiple machines to a single zpoolI have a long-running Debian server (myserver)that’s had everything but the OS moved into a zpool (output of command redacted, file sizes won’t add up): user1@myserver:~$ zfs list NAME USED AVAIL REFER MOUNTPOINT tank 3.02T 505G 104K /tank tank/data 1.50T 505G 6.28M /tank/data tank/data/home 511G 505G 96K /tank/data/home tank/data/home/user1 259G 505G 259G /home/user1 tank/data/home/user2 ...
#zfs
Practical ZFS - Discussions
Migrating a root EXT4 file system to OpenZFS - looking for people who have done it

Edit: original post seemed disjoined…so I redid it.

I’ve been thinking about migrating my root file system, currently on my Debian 12 desktop/server, over to OpenZFS with zfsbootmenu. Snapshots and rollbacks would be incredibly cool and useful. I’ve already been running OpenZFS on my big storage pool for a while, but not on root. Looking back, if I’d been more forward thinking, I would have set this up during the initial Debian 12 install years ago, but alas, I didn’t.

Now, I’m considering how to migrate a root file system (and boot) without completely messing things up. Has anyone here done this and not hated themselves for the hassle? My initial research suggests it’s a pretty risky and heavy lift. I wouldn’t even dream of doing an in place conversion. That’s just crazy talk in my mind. Instead, I’d buy another SSD, create the root and boot pools there, and then attempt to copy everything over.

I’m also wondering if it’s worth the effort. I know that’s subjective. EXT4 isn’t a bad file system, but I can’t create a mirrored pool with it, which is what I’d really like for my root/boot SSDs. My plan is to migrate to a single pool root/boot SSD first, then buy another one to add for mirroring.

Thanks for any experiences or insights you might share if you’ve attempted this!

1 post - 1 participant

Read full topic

Practical ZFS · Migrating a root EXT4 file system to OpenZFS - looking for people who have done itEdit: original post seemed disjoined…so I redid it. I’ve been thinking about migrating my root file system, currently on my Debian 12 desktop/server, over to OpenZFS with zfsbootmenu. Snapshots and rollbacks would be incredibly cool and useful. I’ve already been running OpenZFS on my big storage pool for a while, but not on root. Looking back, if I’d been more forward thinking, I would have set this up during the initial Debian 12 install years ago, but alas, I didn’t. Now, I’m considering how...
#zfs
brett
Just replaced a drive in my backup netbook and started it resilvering :-) Fortunately the hardest part (figuring out which drive was broken) was made trivial because one of the drives had an angry red LED on it.

#ZFS / #backups nerds - should I pause backups for 4 days while it resilvers or can I just let it go?
Practical ZFS - Discussions
Reverse order of snapshots

I’m in the process of migrating a Solaris 11 fileserver to FreeBSD. zfs send wasn’t an option given the divergence in the ZFS implementations and I only copied the latest data because our backup-backup fileserver had died so I was in a hurry. I’m now looking to backfill at least the yearly snapshots. If I hadn’t already transferred the latest data, the sensible approach would have been to work forwards, first transferring the data in the oldest yearly snapshot and then using rsync to go forwards a year at a time creating snapshots as I go. I could now do likewise but go backwards. However, I’d then end up with the snapshots in the wrong order.

Is there any creative way to reverse the ordering? I was first thinking it might be possible with zfs promote but that only transfers dependencies with no ordering inversion. zfs send gives an error if you ask it to create a send stream from a newer to an older snapshot - doesn’t seem very Unixy for it to not simply do as it’s told. Are there any other features I might (ab)use? Scanning the documentation, I find one or two features that I don’t understand well enough to know if they might help such as zfs recv -o origin=.

Reversing changes might be something the zstream command could be adapted to do. Do you know enough about the form of ZFS streams or ZFS in general to say whether that might be impossible, trivial or something in-between? Unless it’s trivial, I’d probably be better-off doing more transfers and wasting the computer’s time rather than mine.

1 post - 1 participant

Read full topic

Practical ZFS · Reverse order of snapshotsI’m in the process of migrating a Solaris 11 fileserver to FreeBSD. zfs send wasn’t an option given the divergence in the ZFS implementations and I only copied the latest data because our backup-backup fileserver had died so I was in a hurry. I’m now looking to backfill at least the yearly snapshots. If I hadn’t already transferred the latest data, the sensible approach would have been to work forwards, first transferring the data in the oldest yearly snapshot and then using rsync to go forwards...
#zfs
Practical ZFS - Discussions
Possible to limit cores used by zfs compression?

Running U24.04. I’m experiencing audio glitching when my filesystem does syncs. I’m running a 9950X with 128GB RAM. When doing somewhat meager things like building a python venv, suddenly all the CPUs light up and it glitches all sound from Firefox. Compression is set to zstd.

I’ve looked through ZFS doesn't respect Linux kernel CPU isolation mechanisms · Issue #8908 · openzfs/zfs · GitHub and that does not appear to have any working advice. Is there anything else I should consider?

2 posts - 2 participants

Read full topic

Practical ZFS · Possible to limit cores used by zfs compression?Running U24.04. I’m experiencing audio glitching when my filesystem does syncs. I’m running a 9950X with 128GB RAM. When doing somewhat meager things like building a python venv, suddenly all the CPUs light up and it glitches all sound from Firefox. Compression is set to zstd. I’ve looked through ZFS doesn't respect Linux kernel CPU isolation mechanisms · Issue #8908 · openzfs/zfs · GitHub and that does not appear to have any working advice. Is there anything else I should consider?
#zfs
Asta [AMP]

Any recommendations for decent used external SAS units? I have literally filled my tower with drives and need moar space. Something used I can get on the cheap is great, something new that isn't awful is also fine. I need to ideally reclaim some space in the unit, for that matter... I can get creative, but in doing so I'll be in "these aren't spots where hard drives were meant to go" territory.

4 to 8 is ideal, I'll never complain about larger.

#SASDrives #RAID #ZFS #NAS #techPosting #techRecommendations

Practical ZFS - Discussions
Multiple Encrypted Pools, overlapping mountpoints, and startup loading

I’m still muddling along in building my desired home NAS/desktop.

I started chasing what I thought was a simple recommendation by Mr. Salter. Putting /var on a non-NVMe pool. I’ve found the rabbit hole to be surprisingly complex. Since ZFS chooses when to load the mountpoint, there seem to be a few ways to do it. Has anyone else done ZFS PAM configurations, ZFS - ArchWiki

This approach also seems to solve how to decrypt all of my encrypted drives. Mount order would allow rootpool/ROOT/ubuntu to mount and then rustpool/var to mount after / , but before the OS tries to do anything.

This was solely to reduce the extensive logging writes to my NVMe pool, but I’m almost ready to pull the rip cord and put /var in a single pool just to ensure I can complete a build.

One of the suggestions will be to incorporate the rust pool into a single overall pool. For now, I want to keep them separate while I’m learning the day to day of ZFS.

All recommendations welcome.

1 post - 1 participant

Read full topic

Practical ZFS · Multiple Encrypted Pools, overlapping mountpoints, and startup loadingI’m still muddling along in building my desired home NAS/desktop. I started chasing what I thought was a simple recommendation by Mr. Salter. Putting /var on a non-NVMe pool. I’ve found the rabbit hole to be surprisingly complex. Since ZFS chooses when to load the mountpoint, there seem to be a few ways to do it. Has anyone else done ZFS PAM configurations, ZFS - ArchWiki This approach also seems to solve how to decrypt all of my encrypted drives. Mount order would allow rootpool/ROOT/ubuntu t...
#zfs
Practical ZFS - Discussions
Automatically import/export pool on startup/shutdown?

Hello, I multi-boot, I have a 3 diskz1 pool with basically my local data that I have been sharing between some Void installations using

/etc/rc.local

#!/bin/sh
# Default rc.local for void; add your custom commands here.
#
# This is run by runit in stage 2 before the services are executed
# (see /etc/runit/2).
zpool import lagoon

&

/etc/rc.shutdown

#!/bin/sh
# Default rc.shutdown for void; add your custom commands here.
#
# This is run by runit in stage 3 after the services are stopped
# (see /etc/runit/3).
zpool export lagoon

How would I do this in systemd distributions? Debian, Ubuntu, Mint, CachyOS etc?

I have found you can start a script with roots crontab, but what about the shutdown-export side?

1 post - 1 participant

Read full topic

Practical ZFS · Automatically import/export pool on startup/shutdown?Hello, I multi-boot, I have a 3 diskz1 pool with basically my local data that I have been sharing between some Void installations using /etc/rc.local #!/bin/sh # Default rc.local for void; add your custom commands here. # # This is run by runit in stage 2 before the services are executed # (see /etc/runit/2). zpool import lagoon & /etc/rc.shutdown #!/bin/sh # Default rc.shutdown for void; add your custom commands here. # # This is run by runit in stage 3 after the services are stopped # (se...
#zfs
AngesagtLive-Feeds
Über