It's probably connected to accessing the sleepimage file during wakeup. No sleep image, no problem
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
MP All Models PCIe SSDs - NVMe & AHCI
- Thread starter MisterAndrew
- WikiPost WikiPost
- Start date
- Sort by reaction score
- Status
It's probably connected to accessing the sleepimage file during wakeup. No sleep image, no problem
Well, success with the 970Pro. I pulled the 970EVO Plus and RMA'ed it back to Newegg. Either a bad blade or some other incompatibility. I didn't really need to spend another $40 for the Pro (512GB version), but I'm glad it's working and booting.
Attachments
So I've tried the new 970 Evo Plus also, and have had more success.
I had to troubleshoot quite a bit too, but it turned out a dated kext was the culprit when CCC'ed over to the blade. I'm running the 500GB variant on a kryoM.2 adaptor, with 141 bootrom and two boot volumes, HS and Mojave. Firmware is 1B2QEXM7. The x4 linkspeed gets fully utilized as expected, and its stable so far. Very pleased with the result.
I had to troubleshoot quite a bit too, but it turned out a dated kext was the culprit when CCC'ed over to the blade. I'm running the 500GB variant on a kryoM.2 adaptor, with 141 bootrom and two boot volumes, HS and Mojave. Firmware is 1B2QEXM7. The x4 linkspeed gets fully utilized as expected, and its stable so far. Very pleased with the result.
Code:
Workstation:~ Fangio$ diskutil list
/dev/disk0 (external):
#: TYPE NAME SIZE IDENTIFIER
0: GUID_partition_scheme 500.1 GB disk0
1: EFI EFI 209.7 MB disk0s1
2: Apple_APFS Container disk7 499.8 GB disk0s2
[...]
/dev/disk7 (synthesized):
#: TYPE NAME SIZE IDENTIFIER
0: APFS Container Scheme - +499.8 GB disk7
Physical Store disk0s2
1: APFS Volume Samsung 970 Mojave 175.0 GB disk7s1
2: APFS Volume Samsung 970 High Sierra 150.0 GB disk7s2
3: APFS Volume Preboot 48.7 MB disk7s3
4: APFS Volume Recovery 1.1 GB disk7s4
5: APFS Volume VM 20.5 KB disk7s5Nope. It's just hard to reproduce. Most of the time it wakes up from a long period of sleep, other times it has this problem.
Can't reproduce it with short sleep/wake cycles.
I guess this SSD is not 100% compatible. OWC Aura Pro X users experienced similar issues and found no solution, theirs wasn't compatible as well.
The adapter card is noname and black, with the two leds (blue and red) but I doubt that's the problem.
The logs from this morning's wake crash:
Code:
com.apple.message.domain: com.apple.sleepwake.failure
com.apple.message.signature3: Wake
com.apple.message.signature: Drivers Failure
com.apple.message.signature2: IOU2(PXS3 IONVMeController),RP03(Intel82574L),RP04(Intel82574L),SATA(AppleAHCI)
com.apple.message.value: 5
com.apple.message.summarize: NO
SenderMachUUID: 82BA3EF4-1A03-3292-A7FA-95B595883B9B
Code:
com.apple.message.domain: com.apple.wake.failure
com.apple.message.signature: Drivers Failure
com.apple.message.signature2: IOU2(PXS3 IONVMeController),RP03(Intel82574L),RP04(Intel82574L),SATA(AppleAHCI)
com.apple.message.summarize: YES
SenderMachUUID: 82BA3EF4-1A03-3292-A7FA-95B595883B9BHS 1.13.6
People who had similar sleep/wake issues with 1.13.2 tried this:
Sound like voodoo with and SSD problem, but hey. They also find it incredibly hard to reproduce. Here's a few months of desperation with various Macs: https://discussions.apple.com/thread/8194934?page=6
Last edited:
OK so is the 970 Evo Plus now confirmed as working and is compatable?
If confirmed is it too early to perhaps update the M.2 NVMe and AHCI Blades info located at start of this thread for the 970 EVO Plus?
Currently there is still no info next to Compatibility status:
970 EVO Plus: Phoenix controller (S4LR020): NVMe
3D TLC (96-layer)
Available in 250GB, 500GB, and 1TB
PCIe 3.0 x4
Speeds: up to 3,500 MB/s read, up to 3,300 MB/s write
Datasheet
Compatibility status:
Last edited:
Just tested a 970 EVO Plus. Had some -83 erros like the user @fatespawn. I used this PCIe adapter: https://www.amazon.de/gp/product/B07GWNB1CX/ref=ppx_yo_dt_b_asin_title_o04_s00?ie=UTF8&psc=1
The 970 EVO plus showed up immediately in MacOS. First attempt to clone failed due the error. After a reboot I gained access to the SSD and the cloning process was successful, no boot was possible afterwards. I put in my old 5770 for bootscreen and started a verbose boot (see attachment).
I even tried a clean install booted from my old SSD, but after the reboot my screen kept black.
I tested the r/w speed of the EVO plus booted up from by old SSD. Read and write speeds decreased within seconds from 1500 to 700. Temperature seems okay with 49 °C, or is it too hot already?
When I'm back at home, I check the firmware / revision of my 970 EVO plus.
I'll keep you updated about this, but I ordered a 970 EVO pro and Lycom DT-120.
The 970 EVO plus showed up immediately in MacOS. First attempt to clone failed due the error. After a reboot I gained access to the SSD and the cloning process was successful, no boot was possible afterwards. I put in my old 5770 for bootscreen and started a verbose boot (see attachment).
I even tried a clean install booted from my old SSD, but after the reboot my screen kept black.
I tested the r/w speed of the EVO plus booted up from by old SSD. Read and write speeds decreased within seconds from 1500 to 700. Temperature seems okay with 49 °C, or is it too hot already?
When I'm back at home, I check the firmware / revision of my 970 EVO plus.
I'll keep you updated about this, but I ordered a 970 EVO pro and Lycom DT-120.
Attachments
Last edited:
As an Amazon Associate, MacRumors earns a commission from qualifying purchases made through links in this post.
In short, if I were you I'd wait for more reports.
I realize this could become a quite popular blade and this may be not exactly what potential buyers want to read here. I will not conceal that during my troubleshooting I've seen some of the KP and Disk Utility First Aid error messages posted by @fatespawn too, also some behaviour from @crjackson2134 's description. I just didn't give up and eventually, I got it to work. Maybe a future firmware update will cure this delicateness.
Anyway this is what First Aid says now (partially translated)
However I'd rather not want people base their decision just on this one positive outcome. I'd like to see some more reports myself before the compatibility status gets changed.
Last edited:
I have not seen a single report of anyone having GOOD luck with the 970EVO PLUS (970EVO+) as a boot drive. No Mac or even Hackintosh, macOS does not seem to like these SSDs as a boot drive. Everyone should use the standard EVO or PRO models if you want a Samsung 970 series SSD
Yes.. this for me too. I forgot to mention that symptom. At one point, I was able to mount the drive and started to clone, but like you I experienced errors and it failed 1/4 of the way through. I did a Blackmagic speed test on it and had the same symptoms - 1500.... then quickly dropping to 800, 700, back up to 1500. I couldn't even get it to work as a data drive, let alone a boot drive. I'm glad @Fangio has had better luck.
Hopefully Samsung may be receiving feedback on the 970 EVO Plus that will allow them to modify the firmware so that it is more Mac compatible.
Looking briefly over the benchmarks it seems that it's got somewhat better sustained write and mixed I/O speed (on PCs) and eventually its price should come down further, but from what I'm seeing I'll probably wind up picking up a 970 1 TB or an ADATA SX8200 Pro to replace the 512 GB 970 Pro in my MP 5,1. Need to upgrade because of multiple (and growing) Windows and Linux VMs.
Looking briefly over the benchmarks it seems that it's got somewhat better sustained write and mixed I/O speed (on PCs) and eventually its price should come down further, but from what I'm seeing I'll probably wind up picking up a 970 1 TB or an ADATA SX8200 Pro to replace the 512 GB 970 Pro in my MP 5,1. Need to upgrade because of multiple (and growing) Windows and Linux VMs.
Thanks for the feedback. I wasn't planning to buy a 970 EVO Plus because there isn't a clear picture about it's compatability. Thought my post might help clarify this, and you guys have clearly detailed your combined thoughts which is exactly what I was hoping for.
I'm running one of the orginal Amfeltec Squid PCI Express Carrier Boards (PCI Express Gen 2 Carrier Board) with two Samsung 512GB SM951 AHCI running as a 1TB RAID 1 Volume, so I'm happy with what I have currently.
I'm running one of the orginal Amfeltec Squid PCI Express Carrier Boards (PCI Express Gen 2 Carrier Board) with two Samsung 512GB SM951 AHCI running as a 1TB RAID 1 Volume, so I'm happy with what I have currently.
I attached a photo of my 970 EVO plus. Maybe @Fangio has another firmware or version of this SSD. Mine was even produced in February, so it should be one of the most recent.
In a few hours I can test my 970 EVO pro, I'll keep you updated.
EDIT: 970 EVO pro did it! Without any problems, r/w stable at 1,500. I can recommend this upgrade. And at the same time i changed from 4x 4 GB RAM to 3x 8 GB. Powerful machine now!
In a few hours I can test my 970 EVO pro, I'll keep you updated.
EDIT: 970 EVO pro did it! Without any problems, r/w stable at 1,500. I can recommend this upgrade. And at the same time i changed from 4x 4 GB RAM to 3x 8 GB. Powerful machine now!
Attachments
Last edited:
Nope. See pic in #628. I can take pics of the box too ofc..
Maybe this will comfirm I wasn't talking BS here
Not sure what else I can do to prove that it works. Maybe send an I/O Registry dump to members knowledgeable enough to read that.
Btw I had to remove my Velocity Solo x2 PCI card for SATA3 first before things started to work.
---
The HS volume has had no problems booting from my CCC clone. In Mojave I've had older versions of Lilu w/ AppleALC, Innie & NightShiftUnlocker as plugins, working fine on the SATA SSD. Pretty sure removing helped to boot from the 970 Evo Plus Mojave volume.
Edit2, worth mentioning again: this 4,1/5,1 has the 141.0.0.0.0 BootROM.
Last edited:
Interesting you say this initiated the resolution. I've noticed some inconsistent behavior with Apricorn Velocity Duo X2. Not enough to say it's not compatible with 140.0.0.0.0 firmware, but enough random issues related directly to this adapter (on occasion only) to be concerned with the adapter moving forward. Was using in conjunction with Velocity Solo X2 before NVMe boot drive (via PX1), but the Solo has since been removed from the system.
Nope. See pic in #628. I can take pics of the box too ofc..
View attachment 825378
Not sure what else I can do to prove that it works. Maybe send an I/O Registry dump to members knowledgeable enough to read that.
The HS volume has had no problems booting from my CCC clone. In Mojave I've had older versions of Lilu w/ AppleALC, Innie & NightShiftUnlocker as plugins, working fine on the SATA SSD. Pretty sure removing helped to boot from the 970 Evo Plus Mojave volume.
Would be very interested to see your KEXT list or something as I've also got the Evo Plus, 1TB with the latest 1B2QEXM7 firmware and I'm having constant kernel panics. Have had to remove it to keep the system stable and obviously, would very much like to be able to use this drive
Well the list of Third Party Kernel Extensions in Mojave is rather short currently
On the other hand, simply kextstat will generate a significant output (meaning, Apple’s official macOS kexts included). Too long to list here, sorry.
I've learned over at macuser.de that Mojave contains a newer version of APFS, so I've initialized the drive there. Adding containers/volumes or delete them on 970 Evo Plus turned out to work better in High Sierra though. I've since moved both volumes at the beginning of the list
Theres a screenshot of how the drives look on the desktop currently. The SATA SSDs are 830s, the gap is where more volumes on an 840 Evo and my movies disk are still missing. I've only got one OWC Mount Pro.
@Siftah Did you file a bug report, and send Samsung some feedback? I did.
Code:
Workstation:~ Fangio$ kextstat | grep -v com.apple
Index Refs Address Size Wired Name (Version) UUID <Linked Against>
82 0 0xffffff7f80e71000 0x17f000 0x17f000 at.obdev.nke.LittleSnitch (5210) 0BEBCAD5-9FC2-35B4-A87D-08B14BCB15F5 <8 6 5 3 1>
95 0 0xffffff7f816c0000 0xb000 0xb000 org.dungeon.driver.SATSMARTDriver (0.8.1) F82FDDF4-8465-52D6-DE01-FA1BE76AB4F3 <29 28 27 5 3>
132 0 0xffffff7f82a72000 0x7000 0x7000 com.hzsystems.terminus.driver (4) A34AA04A-490F-DEE3-BF42-CDAE5B86833D <91 6 5 3 1>On the other hand, simply kextstat will generate a significant output (meaning, Apple’s official macOS kexts included). Too long to list here, sorry.
I've learned over at macuser.de that Mojave contains a newer version of APFS, so I've initialized the drive there. Adding containers/volumes or delete them on 970 Evo Plus turned out to work better in High Sierra though. I've since moved both volumes at the beginning of the list
Code:
Workstation:~ Fangio$ diskutil list
/dev/disk0 (external):
#: TYPE NAME SIZE IDENTIFIER
0: GUID_partition_scheme 500.1 GB disk0
1: EFI EFI 209.7 MB disk0s1
2: Apple_APFS Container disk1 499.8 GB disk0s2
/dev/disk1 (synthesized):
#: TYPE NAME SIZE IDENTIFIER
0: APFS Container Scheme - +499.8 GB disk1
Physical Store disk0s2
1: APFS Volume Samsung 970 Mojave 134.8 GB disk1s1
2: APFS Volume Samsung 970 High Sierra 150.0 GB disk1s2
3: APFS Volume Preboot 70.7 MB disk1s3
4: APFS Volume Recovery 1.1 GB disk1s4
5: APFS Volume VM 20.5 KB disk1s5
/dev/disk2 (internal, physical): [...]
/dev/disk3 (internal, physical): [...]
/dev/disk4 (internal, physical): [...]
/dev/disk5 (internal, physical): [...]
/dev/disk6 (synthesized):
#: TYPE NAME SIZE IDENTIFIER
0: APFS Container Scheme - +127.5 GB disk6
/dev/disk7 (synthesized):
#: TYPE NAME SIZE IDENTIFIER
0: APFS Container Scheme - +255.9 GB disk7
/dev/disk8 (external, physical): [...]Theres a screenshot of how the drives look on the desktop currently. The SATA SSDs are 830s, the gap is where more volumes on an 840 Evo and my movies disk are still missing. I've only got one OWC Mount Pro.
@Siftah Did you file a bug report, and send Samsung some feedback? I did.
Attachments
There are demonstrated / reproducible issues with Samsung 970 NVMe SSD's used in conjunction with the RX580 and AHCI/SATA devices. Switching to a Tahiti based video card, such as the 7950 or 7970, eliminates the issue.
While @Fangio has noted the 970 Evo Plus is working on a cMP with a AMD 280x, for users who are having issues with the Evo Plus.... What video card are you using?
While @Fangio has noted the 970 Evo Plus is working on a cMP with a AMD 280x, for users who are having issues with the Evo Plus.... What video card are you using?
You are making the wrong correlation here, GPU is not the smoking gun. Something else is causing this.
Since lot's of people have 970 PROs working perfectly and the booting problem affect probably two people, maybe we are looking at the wrong thing - I don't doubt that some of 970 PROs could be lemons, but we sure didn't found what is going on yet.
I've been using the Sapphire Pulse RX 580 + 970 Pro in my office Mac Pro since @crjackson2134 started to have problems and I tried to track it unsuccessfully. I boot numerous times a day, shutdown at day's end and I never get this problem.
This mystery continues.
Last edited:
Another data point:
I have a H*ckint*sh with the following parts:
---
8700K
ASUS ROG STRIX-Z390-I motherboard
Sapphire RX 580 Nitro+ SE
Crucial 500GB SATA SSD (The Cx500 or whatever)
970 EVO Plus
Vanilla macOS install with SIP fully enabled.
* NO modifications to the OS install itself.
* Latest Lilu, AppleALC, WhateverGreen versions.
---
Trying to use macOS with the 970 EVO Plus, the system freezes within a minute or two of boot. The screen displays my desktop but it's frozen and the keyboard/mouse don't work and the device doesn't ping from a remote device. So, a hard lock.
So I ended up installing macOS on the Crucial for now, and put Windows on the EVO 970 Plus. The Plus works flawlessly under Windows.
When I boot into macOS off of the Crucial, the 970 EVO Plus with its NTFS volume is automatically mounted. If I do not unmount it quickly the system will freeze within a minute or two in the exact same manner. (I fixed this by turning off auto mount with sudo vifs)
I read the posts in this forum and as a test, removed my RX 580 card from my system completely and the lock up scenarios I described continue to happen just the same.
TL;DR: macOS locks up whenever a partition on the EVO 970 Plus is mounted, whether it's APFS, HFS+, or NTFS. If the partition is not mounted, the system does not freeze. The system otherwise runs remarkably.
Edit: Just to clarify, I'm not looking for help, just offering a data point to this discussion. I'd be happy to answer any questions.
[doublepost=1552156438][/doublepost]
The issue is with the EVO 970 Plus model. The 970 EVO and 970 EVO Pro work fine.
Yes, same here. The Velocity Solo x2 was the main reason to move on to NVMe blades. I've had two 830 SSDs connected to it and had to grab additional power for it from drive bay 4, messy look included. When booted from one of them I can't do any firmware updates in Mojave. It would only work in HS where I did all of them so far (083-089 except 087 thanks to the warnings here, and then 138-141).
Looking forward to try whatever the 10.14.4 release comes up with.