Pool issue

[tehbrd]

Ive run into a strange issue with a zpool running FreeNAS-8.2.0-BETA2-x64 (r10638) which I'm hoping someone can help me with.

It all appeared to start several hours after I made an amendment to /boot/loader.conf after I had a panic (Kernel Panic: Kmem_malloc kmem_map too small). To which I added vm.kmem_size = 1.5x mem and vfs.zfs.arc_max = .5x mem and reloaded. Ive rolled back these changes but I still have an issue with one of my pools. What I now have is when I go to access this particular mounted pool either via a share or straight from the terminal the second level of directories are inaccessible (even a listing freezes). There doesn't appear to be anything of value in syslog either. I've attempted to run a scrub and all this does is freeze the particular session im in and it renders all future status checks against the pool inoperable (ie I can run a status but after attempting a scrub a status freezes the cli). When I attempt to shutdown I also get an error relating to processes that wont die (or something to that effect). My other pool is working fine and its not having any issues.

Im thinking from a bit of reading around perhaps an export/import, but Im interested in seeing if anyone had run into something similar and have a more informed idea.

This is what I see in relation to the pool;

[root@hi] ~# zpool status -x
all pools are healthy

[root@hi] ~# zpool status -v Store
pool: Store
state: ONLINE
scrub: none requested
config:

NAME STATE READ WRITE CKSUM
Store ONLINE 0 0 0
raidz1 ONLINE 0 0 0
ada6p2 ONLINE 0 0 0
ada4p2 ONLINE 0 0 0
ada8p2 ONLINE 0 0 0
raidz1 ONLINE 0 0 0
ada5p2 ONLINE 0 0 0
ada1p2 ONLINE 0 0 0
ada2p2 ONLINE 0 0 0
ada3p2 ONLINE 0 0 0

errors: No known data errors

 

[peterh]

If this is a 64bit cpu AND 64 bit freenas then you should not need any loader tweaks regarding to memory.

You are expected to have at least 8GB of memory however.
If you have turned on dedup you might be screwed ( and need to get much more memory or scratch that filesystem)

 

[tehbrd]

I had opt to tune the memory configuration as I was concerned there was some kind of memory issue despite having 8gb of RAM. But no amount of tuning has seemed to make a difference, as I still get this deadlock condition.

Originally it appeared that there was some kind of issue with a directory/files and once that was accessed that entire pool became locked (one of SABnzbds work directories and it was run on boot), on cleaning up that directory the system now doesn't quick lock down the pool but it now manifest on large data copies and certain par jobs (in relation to SAB).

 

[tehbrd]

I've just triggered the condition trying to remove a file.

load: 0.08 cmd: rm 4533 [zio->io_cv)] 258.66r 0.00u 0.03s 0% 1260k

[root@hi] ~# procstat 4533
PID PPID PGID SID TSID THR LOGIN WCHAN EMUL COMM
4533 3010 4533 3010 3010 1 root zio->io_ FreeBSD ELF64 rm

[root@hi] ~# zpool status -x
all pools are healthy

Seems to be an issue people have come across before for a number of reasons, so more investigation is required!

[root@hi] ~# pciconf -lv | grep -A 1 -B 3 'mass storage'
atapci1@pci0:0:31:2: class=0x01018a card=0xb0021458 chip=0x3a208086 rev=0x00 hdr=0x00
vendor = 'Intel Corporation'
device = 'SATA2(4Port2) (ICH10 Family)'
class = mass storage
subclass = ATA
--
atapci2@pci0:0:31:5: class=0x010185 card=0xb0021458 chip=0x3a268086 rev=0x00 hdr=0x00
vendor = 'Intel Corporation'
device = 'SATA2(2Port2) (ICH10 Family)'
class = mass storage
subclass = ATA
mvs0@pci0:1:0:0: class=0x010400 card=0x02439005 chip=0x02439005 rev=0x02 hdr=0x00
vendor = 'Adaptec Inc'
class = mass storage
subclass = RAID
--
atapci0@pci0:3:5:0: class=0x010185 card=0xb0001458 chip=0x82131283 rev=0x00 hdr=0x00
vendor = 'Integrated Technology Express (ITE) Inc'
device = 'IDE Controller (IT8213F)'
class = mass storage
subclass = ATA