I haven't made any significant configuration changes lately, but I've just started seeing a collection of unusual issues, which seem to be I/O or network-related.

I've seen lines like this in my top output:

14460 root      20   0     0    0    0 R  100  0.0  14:10.88 flush-8:16

I've seen bash hang on tab-completeing, and ls hang, and had to kill (mosh, ssh, telnet) and reconnect. I've seen ^C do nothing. I've seen dd hang trying to write to my cache drive (which passes a reiserfsck). I've seen emhttp fail to respond to requests. Any advice?

Attach a syslog.  For instructions, see this thread:

 

    http://lime-technology.com/forum/index.php?topic=9880.0

Scrubbed some public IPs and private information from log.

SMART report for cache drive: https://gist.github.com/11428370

log.zip

A few `dd`s from /dev/zero to the cache drive:

root@rcombs-NAS:/mnt/cache/.plextmp# dd of=test if=/dev/zero bs=1M
^C6+0 records in
6+0 records out
6291456 bytes (6.3 MB) copied, 12.9394 s, 486 kB/s

root@rcombs-NAS:/mnt/cache/.plextmp# dd of=test if=/dev/zero bs=1M
^C0+0 records in
0+0 records out
0 bytes (0 B) copied, 1.2677e-05 s, 0.0 kB/s

root@rcombs-NAS:/mnt/cache/.plextmp# dd of=test if=/dev/zero bs=1M
^C8+0 records in
8+0 records out
8388608 bytes (8.4 MB) copied, 14.5958 s, 575 kB/s

And then a few from the drive to /dev/zero:

root@rcombs-NAS:/mnt/cache/.plextmp# dd if=test of=/dev/zero bs=1M
8+0 records in
8+0 records out
8388608 bytes (8.4 MB) copied, 0.00148603 s, 5.6 GB/s
root@rcombs-NAS:/mnt/cache/.plextmp# dd if=test of=/dev/zero bs=1M
8+0 records in
8+0 records out
8388608 bytes (8.4 MB) copied, 0.00140512 s, 6.0 GB/s
root@rcombs-NAS:/mnt/cache/.plextmp# dd if=test of=/dev/zero bs=1M
8+0 records in
8+0 records out
8388608 bytes (8.4 MB) copied, 0.0013778 s, 6.1 GB/s

And then I ran

echo 3 > /proc/sys/vm/drop_caches

, and tried those reads again:

root@rcombs-NAS:/mnt/cache/.plextmp# dd if=test of=/dev/zero bs=1M    
8+0 records in
8+0 records out
8388608 bytes (8.4 MB) copied, 13.1475 s, 638 kB/s
root@rcombs-NAS:/mnt/cache/.plextmp# dd if=test of=/dev/zero bs=1M
8+0 records in
8+0 records out
8388608 bytes (8.4 MB) copied, 0.00153729 s, 5.5 GB/s

Those uncached speeds aren't remotely sensible on this SSD, and the SMART report doesn't indicate anything wrong with the actual disk. I'll try updating the firmware later, though.

 

EDIT:

Note that I timed myself when pressing ^C on those writes; they all should've ended 10s after starting, ±2s max.

Start in SAFE-MODE and test again.

Updating my SSD's firmware solved this.