Skip to content

Mobile VOIP Calculation (for PlatinumTel / ptel)

30-Jan-14

Update on 2014-01-29

It looks like I was off by a penny for the VOIP.ms fee. It’s really 1¢/minute, not 2¢/minute. This means that the overall calculation ends up being about 3.23¢/minute, not 4.23¢/minute. This is decently below the 5¢/minute for PlatinumTel. No idea if it performs well (in terms of latency and drop-outs).

Below is the calculation I did for myself around September of last year:

I switched to PlatinumTel on a pay-as-you-go plan. They are a T-Mobile MVNO. I like ’em because they are pretty cheap (5 cents per minute, 2 cents per text), and also they have a pay-as-you-go data option (10 cents a megabyte). The pay-go data isn’t cheap, but we’re usually on WiFi, and it’s just nice to be able to get data if you’re in a pinch. (As far as I know, other pay-go operators don’t allow pay-go data option.)

I’ve been very happy with them. Their coverage is the same as T-Mobile, but (I’m guessing) they don’t support roaming.

From my calcs, the 10c/MB data rate even with an efficient vocoder (G.729) breaks even with the 5c/minute. And the latency isn’t that great. It may end up being cheaper, since I hear (no pun intended) that voice calls have a lot of silence, so you end up being ahead with the VOIP option.

Here are the grueling details of the calculation:
More…

Be the first to like.

Unlike
Filed in Mobile | Tagged , , , , | Comments (0) »

Samba ZFS performance (sequential)

16-Jan-14

ZFS

Be the first to like.

Unlike
Filed in Unix and Linux | Tagged , , , | Comments (0) »

Kingston SV100 64 GB (SV100S264G) Sequential Random Write Benchmark

03-Dec-13

This SSD caught my eye as a ZIL. I wonder how it compares to my previously benchmarked (a number of times) Kingston SV100 64 GB SSD.

I decided to remove it from the ZIL and try it out. I wanted to make sure I exercise the sequential writes (which is all I really care about in a ZIL–I think) with random data, as some SSD controllers get their speed from compression.

In order to do this test, I first wrote the random data to a memory file system. I then wrote from the memory file system to the SSD.

Poojan@server ~ >!df
df -h -t ufs
Filesystem            Size    Used   Avail Capacity  Mounted on
/dev/gpt/usbrootfs     42G    7.4G     31G    19%    /
/dev/md0              247M    8.0k    227M     0%    /var/tmp
Poojan@server ~ >dd if=/dev/random of=/var/tmp/random bs=1M count=100
100+0 records in
100+0 records out
104857600 bytes transferred in 1.048121 secs (100043413 bytes/sec)
Poojan@server ~ >dd if=/dev/random of=/var/tmp/random bs=1M count=100
100+0 records in
100+0 records out
104857600 bytes transferred in 1.048035 secs (100051629 bytes/sec)
Poojan@server ~ >dd if=/var/tmp/random of=/dev/null bs=1M count=100
100+0 records in
100+0 records out
104857600 bytes transferred in 0.012362 secs (8482249780 bytes/sec)
Poojan@server ~ >dd if=/var/tmp/random of=/dev/null bs=1M count=100
100+0 records in
100+0 records out
104857600 bytes transferred in 0.012305 secs (8521529347 bytes/sec)
Poojan@server ~ >dd if=/var/tmp/random of=/dev/null bs=1M count=100
100+0 records in
100+0 records out
104857600 bytes transferred in 0.012351 secs (8489781699 bytes/sec)
Poojan@server ~ >sudo dd if=/var/tmp/random of=/dev/gpt/tank_zil0 bs=1M count=100
100+0 records in
100+0 records out
104857600 bytes transferred in 0.641571 secs (163438858 bytes/sec)
Poojan@server ~ >sudo dd if=/var/tmp/random of=/dev/gpt/tank_zil0 bs=1M count=100
100+0 records in
100+0 records out
104857600 bytes transferred in 0.641162 secs (163543088 bytes/sec)
Poojan@server ~ >sudo dd if=/var/tmp/random of=/dev/gpt/tank_zil0 bs=1M count=100
100+0 records in
100+0 records out
104857600 bytes transferred in 0.641077 secs (163564741 bytes/sec)
Poojan@server ~ >sudo dd if=/var/tmp/random of=/dev/gpt/tank_zil0 bs=10M count=10
10+0 records in
10+0 records out
104857600 bytes transferred in 0.643163 secs (163034263 bytes/sec)

With random data, we’re seeing around 155 MB/s. This SSD is great for sequential writes.

Be the first to like.

Unlike
Filed in Unix and Linux | Tagged , , , , | Comments (0) »

Debugging USB 3.0 Interrupts

28-Oct-13

I’m seeing interrupt storm errors. I’ve also noticed that my USB 3.0 GoFlex 2TB drive (which I have to brag I bought for $67 back in 2009 or so) seems to be intermittently absent from my ZFS pool.

Here’s a vmstat:

root@server ~ >vmstat -i
interrupt total rate
irq1: atkbd0 3163 0
irq16: ehci0 xhci1 4657985391 7471
irq18: xhci0 305433357 489
irq19: atapci0++ 37971283 60
irq23: ehci1 1281657 2
cpu0:timer 3193627253 5122
irq264: ahci0 9319576 14
irq265: hdac0 100 0
irq266: em0:rx 0 165298952 265
irq267: em0:tx 0 29987834 48
irq268: em0:link 9 0
cpu1:timer 483764624 775
cpu7:timer 546385426 876
cpu5:timer 533264630 855
cpu4:timer 577926662 927
cpu3:timer 663537088 1064
cpu6:timer 534384348 857
cpu2:timer 487291455 781
Total 12227462808 19613

Notice the crazy > 7000 interrupt rate on irq16. Note that one downside of vmstat is that it shows the rate over the uptime of the computer–that is, it does no near-term windowing of the data.

So, I decided to switch the port it’s on just to see if it was an issue with the card not liking the drive, or if it was the drive itself. I was surprised that this drive was connected to the built-in Asmedia controller, and not the add-on NEC/Renesas controller. (I would have thought that the built-in controller would have supported MSI rather than falling back to IRQ.)

I’ll check on this tomorrow and see how it’s doing. Probably won’t be scientific, since I don’t know exactly what usage pattern might exercise the interrupts.

Update 1

Oh, it looks like xhci0 is indeed the add-on card:


xhci0: <NEC uPD720200 USB 3.0 controller> mem 0xf7a00000-0xf7a01fff irq 18 at device 0.0 on pci3
xhci0: 32 byte context size.
usbus1 on xhci0
pcib4: <ACPI PCI-PCI bridge> irq 16 at device 28.4 on pci0
pci4: <ACPI PCI bus> on pcib4
xhci1: <XHCI (generic) USB 3.0 controller> mem 0xf7900000-0xf7907fff irq 16 at device 0.0 on pci4
xhci1: 32 byte context size.
usbus2 on xhci1

So, I made have moved the incorrect hard drive (it was the GoFlex that I moved rather than the Backup Plus). However, it is the GoFlex that’s giving checksum errors reported by the pool:


server# zpool status tank
pool: tank
state: ONLINE
status: One or more devices has experienced an unrecoverable error.  An
attempt was made to correct the error.  Applications are unaffected.
action: Determine if the device needs to be replaced, and clear the errors
using ‘zpool clear’ or replace the device with ‘zpool replace’.
see: http://illumos.org/msg/ZFS-8000-9P
scan: scrub repaired 0 in 0h13m with 0 errors on Fri Oct 25 22:31:32 2013
config:

NAME                 STATE     READ WRITE CKSUM
tank                 ONLINE       0     0     0
raidz1-0           ONLINE       0     0     0
gpt/STBPDESK3TB  ONLINE       0     0     0
gpt/WD20EARS     ONLINE       0     0     0
gpt/GOFLEX2TB    ONLINE       0     0     3
mirror-2           ONLINE       0     0     0
gpt/WD15EARS     ONLINE       0     0     0
gpt/ST1500       ONLINE       0     0     0
cache
gpt/tank_cache0    ONLINE       0     0     0

errors: No known data errors

It should also be noted that I’ve been testing a Mushkin USB3.0 drive connected to the xhci0 port, so that might explain the high (average) interrupt rates on xhci0–I’ve been banging away at it for days.

Update 2

Oh: it looks like irq16 is on xhci1, which is the built-in controller, and the port that the GoFlex 2TB drive was attached to. So, this has nothing to do with testing the Mushkin flash drive, and does point to possible failure of the GoFlex drive.

Be the first to like.

Unlike
Filed in Unix and Linux | Tagged , , , , , | Comments (0) »

Home Computer Setup Checklist

27-Oct-13

2017-02-09 Update

As has been widely reported—and a good netizen recently pointed out to me—TrueCrypt is no longer being maintained. I’ve been using Bitlocker To Go (built into Pro versions of Windows) instead. This is only possible on Pro versions of Windows. But, I have Pro where I need it.
There are alternatives. This article has a list.

2015-08-13 Update

Add to the list: Acronis True Image HD, which would basically obviate the need for this list if I made backups often enough—it’s also the only software I’ve successfully used to move data from a hard disk to an SSD (or transfer from disk to disk in general).
Also, add Zint Barcode Studio—I’ve found this quite useful for transferring passwords from computer to phone.
Also, for $25, PDF Creator is now ad-free. Yay!

2014-12-30 Update

Microsoft Security Essentials is a massive fail. Also, I found out that keys between ESET NOD32 Security Suite and NOD32 AV are interchangeable. So, you can buy the security suite and just install it as NOD32 AV (which is much more lightweight).

Original Post

Here’s stuff I install on computers for personal use:

Antivirus: NOD32 AV or Microsoft Windows Security Essentials (MSE).
I got a deal on ESET’s security suite (NOD32 Family Security?), but I found it a bit bulky, and I’d probably pay more NOD32 antivirus, which doesn’t include a lot of the extra functions (such as parental controls, etc.). Usually NewEgg or someone has a good deal on NOD32. I can wait until a good deal on NOD32, and run MSE in the interim (and possibly indefinitely).

I try to install the antivirus before I connect to a network. I download it on another computer–presumably a trusted, known-good, uninfected computer–and transfer it via USB to install it on the new computer.

Web Browser: Firefox
I try to backup my profile directory, so I don’t have to start from scratch each time.

E-Mail: Thunderbird
With Lightning plugin and Provider for Google Calendar.
I try to backup my profile directory, so I don’t have to start from scratch each time.

PDF Reader: Sumatra PDF
Very light-weight, full-feature (at least all the features I want to use) PDF reader.

Disk Encryption: Truecrypt
Although, to be compatible with my (locked-down) computer at work, I’ve been migrating to the built-in encryption of Windows 7 Enterprise/Ultimate.

Office Suite: Microsoft Office (latest edition)
This has gotten a lot more affordable with the home editions.

Password Utility: KeePass2 (AKA KeePass Professional Edition)
with KeePassSync utility (to Amazon S3 storage).

PDF OCR: ABBYY FineReader
This is only on one computer, as I only have 1 license. But that computer is right next to the scanner, so it’s really the only computer than needs it. This was an expensive license (relatively speaking), but it has been well-worth it to be able to search my scanned PDF’s.

Music Management: Media Monkey
I have a license for this, and once again, it’s only installed on one computer. I generally use Subsonic to play music (stream it to Android, iPhone, Firefox, etc). However, Media Monkey allows me to rename my eMusic audibooks with the audiobook name & file number (Part X/Y) in the title. So, rather than each track of “How Music Works” by David Byrne appearing as “Part 1”, “Part 2”, etc., it appears as “How Music Works – Part 1/6”, etc.

Electronic Fax: PamFax
Very reliable, affordable electronic (PDF files etc) to Fax service.

SSH: PuTTY
I run a fewf UNIX/Linux machines, and I used PuTTY to remote login to them.

PDF Writer: PDF Creator or the one that comes with H&R Block (PDF995)?
I dislike the adware bombs that you have to navigate around when you install PDF Creator. This open-source software is good, but the added hassle (and feeling that they are tricking me) has caused me to use the PDF writer that comes with my tax software instead. I’m actually looking for a recommendation here. I don’t mind paying for something, if it is good. In fact, I might pay for PDF creator if they had a premium option that didn’t ask to install some random crapware.—PDF Creator now has an ad-free premium option. Yay!

Version Control (general-purpose): MsysGit & (Optionally) TortoiseGit
TortoiseGit is optional, as sometimes, the command-line is sufficient. In some cases, it’s useful to have a GUI (in the case of TortoiseGit that GUI is integrated into the Windows UI).

Be the first to like.

Unlike
Filed in Desktop | | Comments (0) »

Poojan Wagh Blog

22-Oct-13

pigeons,pigeons I love pigeons!

4 people like this post.

Unlike
Filed in Uncategorized | | Comments (0) »

Kingston SV100S264G 64GB Benchmarks

19-Oct-13

m using a couple of these for log devices (ZIL). I’ve benchmarked them before, but the system is way more stable (got rid of some old SATA cables etc). I removed them from the tank, and here are some dd benchmarks:

They’re roughly in the 150 MB/s range with large (4MB) records, and 32 MB/s with 4kB records.

Be the first to like.

Unlike
Filed in Uncategorized | Tagged , , | Comments (0) »

Mushkin Ventura Pro 64GB USB 3.0 Benchmark

18-Oct-13

As formated out-of-the-box (58.9 GB FAT32):

After hard formatting as exFAT (with default 128kB clusters):

Got it from this slick deal. It actually performs better (on the reads) than advertised (advertised as 120MB/s), and slightly better on the writes (advertised as 70 MB/s).

Update Oct 19

I ran some dd tests on FreeBSD to see what the sector size of this device is. Looks to be around 64kB:

[cce_bash]
server# dd if=/dev/zero of=/dev/da0 bs=16k count=16384
16384+0 records in
16384+0 records out
268435456 bytes transferred in 36.340008 secs (7386775 bytes/sec)
server# dd if=/dev/zero of=/dev/da0 bs=32k count=8192
8192+0 records in
8192+0 records out
268435456 bytes transferred in 18.884614 secs (14214506 bytes/sec)
server# dd if=/dev/zero of=/dev/da0 bs=64k count=4096
4096+0 records in
4096+0 records out
268435456 bytes transferred in 3.608136 secs (74397268 bytes/sec)
server# dd if=/dev/zero of=/dev/da0 bs=128k count=2048
2048+0 records in
2048+0 records out
268435456 bytes transferred in 3.217838 secs (83421065 bytes/sec)
[/cce_bash]

FreeBSD only allows a maximum fragment size of 32kB. Update: I found that you can set a fragment size higher by setting the sector size using the -S option in newfs. If I want to use this as a boot drive, I probably want to enable soft-updates. I’m not crazy about that, as I like to keep every write synchronous. But, the FreeBSD manual says that it’s basically safe–you might lose some data in a 30-second interval, but only if the system crashes–in which case, the time it takes to write to disk on this drive also becomes an issue. Will maybe try that.

This post on FreeBSD on SSD’s recommends soft updates, but more importantly, shows you how to get a swap file that uses TRIM.

Be the first to like.

Unlike
Filed in Desktop | Tagged , , , | Comments (0) »

Plextor 256GB M5 Pro SSD Benchmark

25-Jul-13

Using Crystal Disk Mark and stock firmware (just got it yesterday). Drive label says firmware 1.02. Formated at NTFS with default settings.

———————————————————————–
CrystalDiskMark 3.0.1 x64 (C) 2007-2010 hiyohiyo
Crystal Dew World : http://crystalmark.info/
———————————————————————–
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]

Sequential Read : 515.524 MB/s
Sequential Write : 440.763 MB/s
Random Read 512KB : 395.459 MB/s
Random Write 512KB : 429.231 MB/s
Random Read 4KB (QD=1) : 31.083 MB/s [ 7588.6 IOPS]
Random Write 4KB (QD=1) : 69.657 MB/s [ 17006.0 IOPS]
Random Read 4KB (QD=32) : 152.370 MB/s [ 37199.7 IOPS]
Random Write 4KB (QD=32) : 158.093 MB/s [ 38597.0 IOPS]

Test : 1000 MB [J: 0.0% (0.1/238.3 GB)] (x5)
Date : 2013/07/25 17:13:52
OS : Windows 7 Home Premium Edition SP1 [6.1 Build 7601] (x64)

After upgrade to latest firmware (1.05):

———————————————————————–
CrystalDiskMark 3.0.1 x64 (C) 2007-2010 hiyohiyo
Crystal Dew World : http://crystalmark.info/
———————————————————————–
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]

Sequential Read :   528.983 MB/s
Sequential Write :   449.840 MB/s
Random Read 512KB :   411.411 MB/s
Random Write 512KB :   436.107 MB/s
Random Read 4KB (QD=1) :    30.279 MB/s [  7392.4 IOPS]
Random Write 4KB (QD=1) :    60.589 MB/s [ 14792.2 IOPS]
Random Read 4KB (QD=32) :   221.475 MB/s [ 54071.0 IOPS]
Random Write 4KB (QD=32) :   208.997 MB/s [ 51024.7 IOPS]

Test : 1000 MB [J: 0.0% (0.1/238.3 GB)] (x5)
Date : 2013/07/25 17:41:03
OS : Windows 7 Home Premium Edition SP1 [6.1 Build 7601] (x64)

New firmware greatly improved queue depth of 32 (QD=32) performance.

Be the first to like.

Unlike
Filed in Unix and Linux | Tagged , , , | Comments (0) »

Sandisk 5000 SSD Measurements

16-Dec-12

These are a couple of old 8GB (7.5GbB). Their specs are like so:


server% sudo smartctl -d atacam -i /dev/ada4
smartctl 6.0 2012-10-10 r3643 [FreeBSD 9.0-RELEASE amd64] (local build)
Copyright (C) 2002-12, Bruce Allen, Christian Franke, www.smartmontools.org

=== START OF INFORMATION SECTION ===
Device Model: SanDisk SSD SATA 5000 2.5
Serial Number: 00404300490
LU WWN Device Id: 5 001b47 930303334
Firmware Version: 1.22
User Capacity: 8,012,390,400 bytes [8.01 GB]
Sector Size: 512 bytes logical/physical
Rotation Rate: Solid State Device
Device is: Not in smartctl database [for details use: -P showall]
ATA Version is: ATA/ATAPI-7 T13/1532D revision 4a
SATA Version is: SATA 2.6, 1.5 Gb/s
Local Time is: Sat Dec 15 23:45:03 2012 CST
SMART support is: Ambiguous – ATA IDENTIFY DEVICE words 82-83 don’t show if SMART supported.
SMART support is: Available – device has SMART capability.
SMART support is: Enabled

Doing sequential read measurements (I have two of these SSD’s, one as tank_zil0 and one as tank_zil1):


server% sudo dd if=/dev/gpt/tank_zil1 of=/dev/null bs=1M count=1000
1000+0 records in
1000+0 records out
1048576000 bytes transferred in 14.487853 secs (72376217 bytes/sec)
server% sudo dd if=/dev/gpt/tank_zil0 of=/dev/null bs=1M count=1000
1000+0 records in
1000+0 records out
1048576000 bytes transferred in 14.557015 secs (72032349 bytes/sec)

Doing sequential write measurements:

server% sudo dd if=/dev/zero of=/dev/gpt/tank_zil0 bs=1M count=1000
1000+0 records in
1000+0 records out
1048576000 bytes transferred in 15.465034 secs (67803019 bytes/sec)
server% sudo dd if=/dev/zero of=/dev/gpt/tank_zil1 bs=1M count=1000
1000+0 records in
1000+0 records out
1048576000 bytes transferred in 15.584851 secs (67281747 bytes/sec)

I didn’t want to write directly to the raw device, as I didn’t want to delete the GPT partition table. Anyway, it’s roughly 70 MB/s read and 67 MB/s write.

I was wondering if it’s time to upgrade to a USB 3.0 drive. In particular, the Lexar Triton drive seem really fast. However, they aren’t that much faster–roughly around 110 MB/s in smaller capacities (16GB). As of time of writing, they’re available on NewEgg for around $40. That would set me back $80 for 2, albeit at double the capacity of my current ZIL. (I have my ZIL mirrored).

The SanDisk Extreme 16GB comes close; users report around 58 MB/s sequential (1MB block size) writes, and it currently goes for around $25. Similarly, the ADATA S102 in 32GB capacity is around 50 MB/s write (according to manufacturer specs).

My point is that I bought the SanDisk SATA SSD’s years ago, and they are still quite fast. I don’t feel a need to upgrade. I was contemplating upgrading since I am running short on SATA ports. However, I will probably buy a (cheaper) USB 3.0 enclosure and move a hard drive outside. (The additional hard drives are merely for backup.)

Be the first to like.

Unlike
Filed in Unix and Linux | Tagged , , | Comments (0) »