Take the 2-minute tour ×
Ask Ubuntu is a question and answer site for Ubuntu users and developers. It's 100% free, no registration required.

I'm using a SSD for about 460GB of data on an Ubuntu 12.04 server. (The drive holds 560GB, so it is about 88% full.) I have had a program which does a lot of random access to the data (no writing), which I last ran (normally) a few days ago. When I ran it yesterday, it was suddenly significantly slower than before. The program just does random access to the drive.

Previously I could do about 5000 random lookups per second; now I'm only getting about 100. This is slower than doing random access to a HDD. (I benchmarked a few problems last year.)

The only thing I've done recently was to install gcc 4.7 and to upgrade all my packages. But, I tried compiling everything with clang and didn't see a difference.

The drive is formatted as ext4 with the only options being errors=remount-ro. I tried rebooting the machine and trimming the device but it didn't change things. Profiling the code shows it is spending all its time in open, close and mmap calls. (Note that I'm not calling mmap directly -- I'm using all C-style fopen fseek and fread calls.)

Any idea what could cause this? I could re-format the drive if there was a chance it would help.

Edit: here is some benchmark data on a 2TB HDD and the 500GB SSD

sudo hdparm -Tt /dev/sdb1

/dev/sdb1:
 Timing cached reads:   6814 MB in  2.00 seconds = 3410.05 MB/sec
 Timing buffered disk reads: 458 MB in  3.00 seconds = 152.45 MB/sec

sudo hdparm -Tt /dev/sdc1

/dev/sdc1:
 Timing cached reads:   6890 MB in  2.00 seconds = 3447.93 MB/sec
 Timing buffered disk reads: 780 MB in  3.01 seconds = 259.36 MB/sec

Here is the output from smartctl:

=== START OF INFORMATION SECTION ===
Model Family:     Intel 320 Series SSDs
Device Model:     INTEL SSDSA2CW600G3
Serial Number:    CVPR140004B7600FGN
LU WWN Device Id: 5 001517 9596df196
Firmware Version: 4PC10362
User Capacity:    600,127,266,816 bytes [600 GB]
Sector Size:      512 bytes logical/physical
Device is:        In smartctl database [for details use: -P show]
ATA Version is:   8
ATA Standard is:  ATA-8-ACS revision 4
Local Time is:    Sun Jan 26 16:46:53 2014 MST
SMART support is: Available - device has SMART capability.
SMART support is: Enabled

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED

General SMART Values:
Offline data collection status:  (0x00) Offline data collection activity
                    was never started.
                    Auto Offline Data Collection: Disabled.
Self-test execution status:      (   0) The previous self-test routine completed
                    without error or no self-test has ever 
                    been run.
Total time to complete Offline 
data collection:        (    1) seconds.
Offline data collection
capabilities:            (0x75) SMART execute Offline immediate.
                    No Auto Offline data collection support.
                    Abort Offline collection upon new
                    command.
                    No Offline surface scan supported.
                    Self-test supported.
                    Conveyance Self-test supported.
                    Selective Self-test supported.
SMART capabilities:            (0x0003) Saves SMART data before entering
                    power-saving mode.
                    Supports SMART auto save timer.
Error logging capability:        (0x01) Error logging supported.
                    General Purpose Logging supported.
Short self-test routine 
recommended polling time:    (   1) minutes.
Extended self-test routine
recommended polling time:    (   1) minutes.
Conveyance self-test routine
recommended polling time:    (   1) minutes.
SCT capabilities:          (0x003d) SCT Status supported.
                    SCT Error Recovery Control supported.
                    SCT Feature Control supported.
                    SCT Data Table supported.

SMART Attributes Data Structure revision number: 5
Vendor Specific SMART Attributes with Thresholds:
ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED  WHEN_FAILED RAW_VALUE
  3 Spin_Up_Time            0x0020   100   100   000    Old_age   Offline      -       0
  4 Start_Stop_Count        0x0030   100   100   000    Old_age   Offline      -       0
  5 Reallocated_Sector_Ct   0x0032   100   100   000    Old_age   Always       -       0
  9 Power_On_Hours          0x0032   100   100   000    Old_age   Always       -       17833
 12 Power_Cycle_Count       0x0032   100   100   000    Old_age   Always       -       13
170 Reserve_Block_Count     0x0033   100   100   010    Pre-fail  Always       -       0
171 Program_Fail_Count      0x0032   100   100   000    Old_age   Always       -       0
172 Erase_Fail_Count        0x0032   100   100   000    Old_age   Always       -       0
183 Runtime_Bad_Block       0x0030   100   100   000    Old_age   Offline      -       0
184 End-to-End_Error        0x0032   100   100   090    Old_age   Always       -       0
187 Reported_Uncorrect      0x0032   100   100   000    Old_age   Always       -       0
192 Unsafe_Shutdown_Count   0x0032   100   100   000    Old_age   Always       -       5
199 UDMA_CRC_Error_Count    0x0030   100   100   000    Old_age   Offline      -       0
225 Host_Writes_32MiB       0x0032   100   100   000    Old_age   Always       -       3236658
226 Workld_Media_Wear_Indic 0x0032   100   100   000    Old_age   Always       -       4973
227 Workld_Host_Reads_Perc  0x0032   100   100   000    Old_age   Always       -       65
228 Workload_Minutes        0x0032   100   100   000    Old_age   Always       -       1067143
232 Available_Reservd_Space 0x0033   100   100   010    Pre-fail  Always       -       0
233 Media_Wearout_Indicator 0x0032   096   096   000    Old_age   Always       -       0
241 Host_Writes_32MiB       0x0032   100   100   000    Old_age   Always       -       3236658
242 Host_Reads_32MiB        0x0032   100   100   000    Old_age   Always       -       6437718

SMART Error Log Version: 1
No Errors Logged

SMART Self-test log structure revision number 1
No self-tests have been logged.  [To run self-tests, use: smartctl -t]


Note: selective self-test log revision number (0) not 1 implies that no selective self-test has ever been run
SMART Selective self-test log data structure revision number 0
Note: revision number not 1 implies that no selective self-test has ever been run
 SPAN  MIN_LBA  MAX_LBA  CURRENT_TEST_STATUS
    1        0        0  Not_testing
    2        0        0  Not_testing
    3        0        0  Not_testing
    4        0        0  Not_testing
    5        0        0  Not_testing
Selective self-test flags (0x0):
  After scanning selected spans, do NOT read-scan remainder of disk.
If Selective self-test is pending on power-up, resume after 0 minute delay.

[Edit: I'm continuing to work on this in the background. When I do resolve it, I will post something here.]

share|improve this question
    
First benchmark your ssd.then have a look at its spin time.If its low,increase it(but note that this will wear it out early). Then have a benchmark again to see if anything differs. –  Registered User Jan 26 at 6:56
    
@adityapatil I'm not familiar with how to look at/set the spin time. I've put benchmark specs into the question. –  Nathan S. Jan 26 at 7:11
    
@NathanS. sdc is the SSD? What SSD model is this? The I/O operations on the HDD didn't slow down? Please also provide the output of ˙sudo smartctl -a /dev/sdc`. (If you don't have that program first install with sudo apt-get install smartmontools.) This looks like strange situation, hopefully it is not some kind of hardware problem. I would try powering down the machine totally, unplug the SSD from the power, wait a few minutes and then replug it an test again. (if you have physical access to it) –  falconer Jan 26 at 11:21
    
@NathanS. There is also a minimal chance that, since your SSD is almost full with a lot of static data, the wear leveling algorithm has just kicked in and the SSD is rearranging the data on the drive surface to level the wear on the cells. With a proper firmware this shouldn't cause such a slow-down in user experience, but who knows. If this unlikely case is your problem, then this problem will go away on its own after the SSD has finished the wear leveling process, which should happen in a few 10 minutes, just leave it do its job. Also you have a SATA2 or SATA3 controller? –  falconer Jan 26 at 11:27
    
@falconer The SSD is sdc; it is an Intel 320 series. I don't believe the HDD has slowed down -- I can check this in a bit. I've edited my post with more information, although I'm not sure whether the controller is SATA2 or SATA3. –  Nathan S. Jan 26 at 23:49
show 1 more comment

1 Answer 1

up vote 2 down vote accepted

So, the final answer is a bit embarrassing, as I should have been able to figure it out, but I was looking in the wrong place. The crucial clue was in the question regarding the speed of the HDD -- tests indicated that it had slowed down as well.

My class that handles reading the data on disk has a default read size for its internal buffers. When doing random access this should be relatively small, because you aren't going to use any of the buffered data later. When performing operations across the full file sequentially, larger values give better performance because you will re-use the data. Given that this was a constant in a piece of shared code, tuning for one application using the code essentially broke the tuning for random access, thus causing the massive slowdown I was experiencing.

share|improve this answer
    
you can do some tweak and optimize for your ssd to make it faster namhuy.net/1563/… –  Blanca Higgins Jun 7 at 17:32
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.