bug

[mtiburs]

Dear developers nas4free,

(This is a translation, I do not speak English)

Through this message, I want you to go up on a bug webgui of synchronization on the configuration datasets with identical names.

Here's an example:

nas4free: ~# zpool create raidpool /dev/ada1
nas4free: ~# zfs create raidpool/media
nas4free: ~# zpool create wd60efrx /dev/ada2
nas4free: ~# zfs create wd60efrx/media
nas4free: ~# zfs list
NAME USED AVAIL REFER MOUNTPOINT
raidpool 372K 464M 96K /raidpool
raidpool/media 96K 464M 96K /raidpool/media
wd60efrx 480K 208M 96K /wd60efrx
wd60efrx/media 96K 208M 96K /wd60efrx/media

There, if we: "Disks|ZFS|Configuration|Synchronize"
the webgui will "crush" each name that is identical datasets, only the last datasets will be considered.
This is a bug, it had already been raised, but never got an answer:
viewtopic.php?f=66&t=7925
One year later, a user has posted on the following forum post:
viewtopic.php?f=35&t=10119

I tried to find a troubleshooting solution to have a backup solution, making this way:
- create each datasets with different names
- generate the config for the web interface
- retrieve the config file
- manually rename the config ZFS command line and in the config file
- restore the configuration manually changed

whether:
nas4free: ~# zfs rename wd60efrx/media wd60efrx/totomedia
in "Disks|ZFS|Configuration|Synchronize", to sync
then, make a backup in "System|Backup/Restore"

Edit this file .xml (kind: config-nas4free.local-20151217141242.xml)
and there, making imperative the following two operations
1) search for the zone "datasets":
<datasets>
<dataset>
<uuid>7c720fd9-224d-4ae9-a378-4eb27eee76ef</uuid>
<name>totomedia</name>
<pool>wd60efrx</pool>
...
And change the name of the dataset "totomedia" in media
the aim being to obtain:
<datasets>
<dataset>
<uuid>7c720fd9-224d-4ae9-a378-4eb27eee76ef</uuid>
<name>media</name>
<pool>wd60efrx</pool>
...

then,
2) change the dataset:
nas4free: ~# zfs rename wd60efrx/totomedia wd60efrx/media

That is, we put the names as they should be (we had changed the names for the webgui gives us the "uuid" datasets.

Finally, we must restore the config that has been changed, restart the server and everything will be in order.

There, the dataset "in ZFS" will be "in tune" with "config webgui", but do not do sync, otherwise re-generate the bug and lose our elements, and therefore, repeat this type of operation.

The problem is identical with ZFS volume sets.

But ... the best would be that the bug no longer exists

[JoseMR]

Hello, I used to have similar issues while importing and syncing ZFS pools/datasets through the Web GUI, only one pool is actually imported/considered , and the rest ignored till a reboot, noticed primarily on manually GPT partitioned/formatted drives.

My solution is simply executing through Advanced|Execute command "zpool import -f -a" and after import then syncing the pools through the Web GUI.

On the datasets with the same names, only the last pool/dataset is actually imported, and is a ZFS bug/limitation, not NAS4Free, but IMHO having name duplicates in ZFS pools/datasets is abit weird also a pain to manage/administer.

Regards

[mtiburs]

I left on consignment (in response to a post) not create the same names datasets

instead of:
pool1/foo
pool2/foo

to a naming style:
pool1/pool1_toto
pool2/pool2_toto

Finally, no matter the manner, the whole is not to create Datasets or volsets of the same name

Regards