With TRIM, Errors - Without TRIM, No Errors??

[erice]

Hello,
in High Sierra , after activating the trim, the boot stays about 20 seconds before these rows:

kernel: (apfs) er_state_obj_get_for_recovery:3682: No ER state object - rolling is not happening, nothing to recover.

kernel: (apfs) handle_mount:254: vol-uuid: 42581E64-0633-4338-A556-AFBEFF7E020F block size: 4096 block count: 58556556 (unencrypted; flags: 0x1; features: 1.0.2)

kernel: (apfs) apfs_vfsop_mount:1331: mounted volume: VM

kernel: (apfs) apfs_load_inode:3805: *** reset ino 775206 size back to 29314 (from 29989)

Instead, without trim, no breaks.

A person of good will might explain to me why?

Sorry for my english and thanks in advance.

 

[dirtydignity]

Did you ever find the answer to this? I am facing the same issue.

 

[g4fa]

Hello,
in High Sierra , after activating the trim, the boot stays about 20 seconds before these rows:

kernel: (apfs) er_state_obj_get_for_recovery:3682: No ER state object - rolling is not happening, nothing to recover.

kernel: (apfs) handle_mount:254: vol-uuid: 42581E64-0633-4338-A556-AFBEFF7E020F block size: 4096 block count: 58556556 (unencrypted; flags: 0x1; features: 1.0.2)

kernel: (apfs) apfs_vfsop_mount:1331: mounted volume: VM

kernel: (apfs) apfs_load_inode:3805: *** reset ino 775206 size back to 29314 (from 29989)

Instead, without trim, no breaks.

A person of good will might explain to me why?

Sorry for my english and thanks in advance.

Same problem and no solution for me
What kind of SSD have you ? is its fresh install or update Sierra ?

 

[GeraldB]

I have the same issue on my build since update to High Sierra. Everything work fine but this message at boot is killing me. So, TRIM disable will disable that message? I have 2 Samsung EVO 850 SSD and 1 old harddisk in my case.
At start it shows:
APFS start 1384
efi_fusion_pairing . something not rolling message.
efi_fusion_pairing . something not rolling message.

B.t.w. For me it doesn't change when I turn off Trim. I have tried that. Also removing other drivers at startup doesn't change a thing. I don't have a fusion drive or never have used one.

 

[svencux]

i am not an expert, but my understanding was that with the new APFS disk format
used in high sierra trim is nativley implemented.

When i do activate the Trim kernel patch in clover, my system boots very slow.

http://www.zdnet.com/article/wwdc-2017-macoss-new-file-system/

 

[vulgo]

I have the same issue on my build since update to High Sierra. Everything work fine but this message at boot is killing me. So, TRIM disable will disable that message? I have 2 Samsung EVO 850 SSD and 1 old harddisk in my case.
At start it shows:
APFS start 1384
efi_fusion_pairing . something not rolling message.
efi_fusion_pairing . something not rolling message.

Attached apfs boot messages with trim patch enabled/disabled. Time to login screen is noticeably longer with the patch enabled (5<10 sec). The messages look similar to me.

 

[GeraldB]

Attached apfs boot messages with trim patch enabled/disabled. Time to login screen is noticeably longer with the patch enabled (5<10 sec). The messages look similar to me.

On my Hack the display of this message is for about 5 seconds. B.t.w. I don't use a trim patch. Only use "sudo trimforce enable".

 

[vulgo]

On my Hack the display of this message is for about 5 seconds. B.t.w. I don't use a trim patch. Only use "sudo trimforce enable".

In the file i posted in #6 there's ~8 second delay, between 12:18:05

(apfs) apfs_mount_update:17785: er: mount check: ro->rw update: no encryption rolling in progress, bailing.

and the next message from apfs at 12:18:13

kernel: (apfs) er_state_obj_get_for_recovery:3682: No ER state object - rolling is not happening, nothing to recover.

which isn't there without trim (the 8 second delay). The output isn't that useful as the only messages for the window relate to encryption rolling (on an unencrypted disk) - BTW i've heard Apple users saying use of trimforce can have a similar effect on boot time.

As my SSD seems to have no problems with Apple's trim implementation i'm leaving the patch enabled for now.

 

[UtterDisbelief]

In the file i posted in #6 there's ~8 second delay, between 12:18:05

(apfs) apfs_mount_update:17785: er: mount check: ro->rw update: no encryption rolling in progress, bailing.

and the next message from apfs at 12:18:13

kernel: (apfs) er_state_obj_get_for_recovery:3682: No ER state object - rolling is not happening, nothing to recover.

which isn't there without trim (the 8 second delay). The output isn't that useful as the only messages for the window relate to encryption rolling (on an unencrypted disk) - BTW i've heard Apple users saying use of trimforce can have a similar effect on boot time.

Hi @vulgo And @GeraldB

I now wonder about trimforce enable. Clearly macOS has the facility to reuse data blocks before waiting for them to be emptied. The trimforce command is built-in, so logically must the software. Some manufacturers have their own hardware routines which do the clean-up while the SSD is powered but not being read.

If macOS doesn't enable trim, is this because the SSD is doing its own cleaning, or because we aren't using a 'genuine' Apple SSD? Should we now leave it be?

I was tempted to disable trimforce again, but reading the Crucial tech docs for my drive, their routines run alongside any native trim, helping rather than hindering. Maybe this is different for different brands?

 

[vulgo]

If macOS doesn't enable trim, is this because the SSD is doing its own cleaning, or because we aren't using a 'genuine' Apple SSD? Should we now leave it be?

The latter. Apple disables features for third party SSDs in IOAHCIBlockStorage because the firmware may be buggy. Linux does something similar.