0.9.9 rfxcv.sys bluescreen Topic is solved

[mattswest]

This crash is occurring on my storage server which runs Windows Server 2012 R2. This most recent time it occurred during a large network file transfer to the machine.

My setup uses a 160GB conventional laptop hard drive as the system drive (C:), a Storage Spaces pool consisting of 3x3TB WD Red NAS drives for storage (E:), and a 32GB SSD used for L2 caching.

My PrimoCache setup is as follows:

Cache Task #1
E: NTFS 5585.87GB 4096.81GB 4k
Status Active
Level-1 Cache 8192MB (MM)
Level-2 Cache 20477MB
Block Size 32KB
Strategy Write-data Only
Defer-Write 30s

Cache Task #2
C: NTFS 148.53GB 117.72GB 4k
Status Active
Level-1 Cache 256MB (MM)
Level-2 Cache 10047MB
Block Size 128KB
Strategy Write-data Only
Defer-Write 10s

I make frequent use of my storage server and such crashes are relatively infrequent. I also use PrimoCache on two other computers without issue, although neither of those use a L2 Cache.

I will email the relevant minidump files to [email protected] after I post this thread.

*Edit: I have sent the minidump files with the subject line "0.9.9 rfxcv.sys crash minidump files (2465)"

[Bjameson]

I removed my previous post because I now also get rxfcv.sys BSODs. The first one happened a few days ago while uninstalling a program from a VM. I dismissed it as a freak incident. But just 5 minutes ago another VM crashed, also while deleting a program. Both VMs are VMWare Player clients running Windows 8.1-X64. The first time L2 was full, the second time half-full.

1 GB RAM R/W cache, 300 secs write delay. 16 GB L2 on a known good SSD. It leaves no crash dump. Instead it shows the message SYSTEM_THREAD_EXCEPTION_NOT_HANDLED (rxfcv.sys). After that it reboots and I have minor corruption on the cached disk (2TB spinner).

I'll now start experimenting with different cache settings. For now, I'll use read-only until I figure out what the problem is.

[CrypEd]

The problem actually might be that your Virtualisation software starts special CPU instruction to process data it should have written to disk. Defer write might interfer virtualisation. Also I do not suggest you to use 5 Minute defer-write timeout with only 1GB R/W-L1-cache...especially when you need the L1-Cache part to load of to the L2.

Please put it to 10 seconds and see if that helps. Deatcivate the seprate caches and check individually. Than we know which configuration is causing th eproblem. I guess the least.

Your operating-drive C: is with with 256MB ... that might be too low, according to first post. Power failure always have risk to create minor-corruptions and ntfs-errors, no matter what the root-cause of the malfuction was, you cannot tell from the corruption on your storage, that this is the causing part.

1 GB RAM R/W cache, 300 secs write delay. 16 GB L2 on a known good SSD. It leaves no crash dump. Instead it shows the message SYSTEM_THREAD_EXCEPTION_NOT_HANDLED (rxfcv.sys). After that it reboots and I have minor corruption on the cached disk (2TB spinner).

I don't get it, this reads totally different than the setup in your initial post. 1GB R/W-Cache with 300secs to small and too high write-timeout.

[Bjameson]

Thanks CrypEd. I'm constantly changing my cache configuration so indeed my posts may seem inconsistent.
The minor corruption I was talking about was easily fixed by Windows. Nothing critical was lost.
I already suspected that the VM process was causing the problem. VMs are always a bit tricky. If the VM is indeed the source of the problem, so be it. I can live with that and I'll avoid certain cache configurations. So far using only L1 R/W 8 GB 300 sec. delay appears completely stable with VMWare, Virtualbox and Hyper-V.
I'll try your suggestions and see what happens.
Thanks.

[Turi]

Hi, I have same problem but without virtualisation software.

Recently installed SSD SATA for use L2 cache on only free port PC, through eSATA.

When reboot PC the L2 cache losses (v0.9.2)/corrupted (v0.9.9) and blue screen.

The minor corruption I was talking about was easily fixed by Windows. Nothing critical was lost.

In my case, lot of corruption and reinstall/repair W2003 server, profile saved and it was not so bad.

¿when stable version?

Thanks.

[Support]

This issue is fixed in version 1.0.1. Thanks all!