You are not logged in.
According to googled reports, no Linux driver is available for ICH7R intel raid.
As I understand that statement, it applies to the raid controllers in ICH7R chipsets which are Windows compatible.
If one has an Intel ICH7 chipset, only two raid hdd's can be utilized as I understand the situation. However, it isn't clear whether that applies to "hardware raid" or "software raid".
Since my mobo has four sata interfaces ICH7 chipset, I am assuming with "software raid" with mdadm that I can make two raid0 arrays or four sata devices in one raid0 array.
However, the Intel system has power sequencing provided for the raid systems to prevent overloads on the power supply should all sata hdd's start at once. I assume this is characteristic of ICH7R chipset( and all other -R chipsets).
I have a system which utilizes sata to Compact Flash adapters, thus the turn-on transient is negligible, since the units are not sata hdd"s. I assume that the sata controller still applies the power sequentially, (in ICH7R only as opposed to ICH7). The interface produces 82MB/s reads with UDMA Compact Flash in mdadm setup.
One statement about raid says that the intel ICH7R is required for raid systems using 3 or more sata devices. Is this the result of power transient consideration?
Is it possible with mdadm to run all four sata interfaces in raid0 with an IntelICH7 chipset if one uses sata to compact flash adapters to eliminate the transient power surge on all four sata interfaces?
Are there any other considerations which limit the ICH7 chipset to two sata devices in raid0?
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
I wouldn't be worried by power surge caused by 4 hdd starting. Forget the mobo driver-raid and go for mdadm. For 4 storage devices you could make raid0, 0+1,1+0 or 5. Depends how much fault tolerance you want.
Offline
Obi-lan;
Thanks for reply....
My concern was that the googled data indicated that ICH7 was limited to two drives ....to utilize three or more requires ICH7R in the mobo according to the data provided.. My mobo is ICH7 only.
The description is not clear as to whether the ICH7 mobo will provide more than two drives if using mdadm in fakeraid(software raid).
Much confusion (FUD) appears in the web concerning raid setups. Raid via pcie is poorly outlined and most results for any raid arrangement do not well describe the read/write speeds now available.
Since I have sata to Compact Flash devices , I get 82MB/sec with two drives. There are 4 sata interfaces available but whether four can be utilized with the ICH7 mobo is not at all clear. The setup with sata recognizes the Compact Flash as UDMA4 devices...not full SATAII speed...limited by compact flash UDMA4 to 42MB/s.
I am not using sata HDD"s obviously...my mobo has only three sata to Compact Flash devices and is a "Live" system with Faunos as the operating system contained on one such sata-cf adapter(42MB/s read).. The raid array is used to provide a fast read-only install of repository packages. With four 8GB UDMA CF cards I would be able to read at 160+MB/s if the mobo can support four such sata raid devices in mdadm format. Not certain at this point whether ICH7R is a hardware raid or a fakeraid.
This setup provides a faster boot since most of the packages are in repo and are accessed with bookmarks in root shell as desired.
Mobo is gigabyte-945gcm/s2c with Pentium4 core 2 duo @ 2.8GHZ with 4 SATAII ports (orange).
I am hoping that the new CFAST compact devices will interface with SATAII with appropriate adapters and thusly provide a possible 300MB/sec from a single CFAST flash device.
The FSB needed to support such speed probably requires a better mobo...not entirely sure on this point. It may max out at 266MB/s with a different CPU on this mobo.. That is an acceptable speed IMHO. A raid0 array of two CFAST would double the capacity @ the same speed.
Anyhow, CFAST is at least a year away, unless cards are provided before the Cameras are upgraded..which is likely to enable testing and integration.
A CFAST to SATA adapter for the Faunos boot would also aid in reducing boot time. The expectation IMHO is that the adapter would probably reduce the actual speed from the advertised 300MB/s to some unknown lesser value.
And there is SATAIII..........................?
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
Never seen mobo with hardware raid. Never seen mobo with unusable connectors. I think you just need to disable all raid stuff in bios, plug in 4 flash-sata devices and check with fdisk etc if linux sees all 4 devices. If it does, then you are good to go with mdadm. Mdadm doesn't work directly with any hardware devices, it just works any harddisk partitions what system can see.
And yeah, sata3 is coming but I think it's only a draft currently. It has 6gbit/s bandwith. I wonder if SAS can outperform fibre storages after this.
Offline
You are probably right about the hardware raid on mobos. It seems it is available on pcie add on cards for a good high price.
I will need to get another adapter from addonics and another pair of 8GB UDMA CF cards to enable the system to repo 32GB and read at 168MB/s, a good clip. That would read a kde repo file in two secs and install in another 5 secs.
I am assuming that Cfast with the sata adapter (not yet designed!) will outperform any sata drive in a sataII system such as mine. I base this on the performance of the present UDMA limited CF cards with Sata adapter which provides no less than predicted speed from the combo. SataII drives do not reach 300MB/sec sustained speeds AFAIK.
Thanks for the comments.....................
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
I think WB Velociraptor 10krpm hdd read/write performance is somewher around 100MB/s both ways. Dunno for faster 15krpm SAS drivers.
Offline
My system does not use any sata hdd drives..they are replaced by sata to Compact Flash devices.
I have three such devices in my system and use USB for any additional devices. Thus the mobo has only three compact flash units connected to sata ports and no pci devices nor agp devices...practically a bare mobo!
I am hoping that the Cfast devices will provide the sata promised 300MB/sec or thereabouts.
My mobo draws 90watts as booted.
The FaunOS system runs entirely in 400mhz ram and is fast!
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
I note the SSD drives can reach ~160MB/sec which may be indicative of the expectation for Cfast which may have different interface specs.
The difference between single layer cells and multi-layer cells is significant in SSD drives, the slc being the more desired speed-wise.
It also seems that smaller capacity units may perform more speedily.
The mobo controller for sata11 also is a question mark as to Cfast performance when it finally gets 'on the street".
Should a 16GB Cfast device reach 160MB/s, it would be a great device for raid0!
Maybe by Xmas?
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
Recently installed a third sata to compact flash device and assembled a three device raid0 array. This was successful with three 8GB UDMA capable Compact Flash devices.
The speed of two such sata/CF devices was previously reported as 82MB/sec in raid0.
The result with hdparm -t in the three device setup was disappointing at 67MB/sec.
This speed was verified in a video transfer of 669mb video which took 10 secs to transfer.
The results indicate that the parameters for a three device raid0 may need a different setup than was provided for the raid0 with two devices.
If changing the params doesn't help, the use of three raid0 devices seems inappropriate.
My mobo is pentium4 on a ga945gcm/S2 with four sata ports...no pci devices...no HDD's.................
So goes the war!
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
Performed the experiment again and received the identical result.
This indicates that the mobo limits the raid0 to two master drives with ich7 as opposed to ich7r which is supposed to enable 3 raid devices in raid0.
Last edited by lilsirecho (2008-10-22 17:36:59)
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
Actually, three drives were registered in the experiment as reported by mdadm -D but suspect that the bus freq of 100mhz for IDE operations(IDE as required by the CF UDMA mode)limits the performance of a three drive raid0 to read speeds less than a two-drive raid0 setup. The IDE bus freq in this mobo is 100mhz.
Why this limiting occurs is not fully apparent to this user except for the added complexity which certainly has the potential to affect the performance.
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
Added a pci card providing two IDE channels with an on-board bios chip and connected an addonics CF/IDE adapter with an 80 wire cable to a third 8GB CF card.
This arrangement provided an hda identifier for the CF card when booted.
Thus, a third CF card was made available for a 3 device raid0 array.
In mdadm the three device raid0 was successfully created. Hdparm reported 93MB/sec for the three device raid0. This was confirmed by a 1GB transfer of video files at the sustained rate of 92MB/sec. A two device CF system in raid0 gave 82MB/sec.
The increase in read speed was not as great as desired, however, there is a limitation applied to the hda CF card because the bus speed for pci is 33mhz.
The experiment shows that the devices can be mixed between hd and sd devices.
[root@n6re ~]# cat /proc/mdstat
Personalities : [raid0]
md1 : active raid0 hda[0] sdb[2] sda[1]
23837376 blocks 64k chunks
unused devices: <none>
[root@n6re ~]# mdadm -D /dev/md1
/dev/md1:
Version : 00.90.03
Creation Time : Tue Oct 28 18:08:48 2008
Raid Level : raid0
Array Size : 23837376 (22.73 GiB 24.41 GB)
Raid Devices : 3
Total Devices : 3
Preferred Minor : 1
Persistence : Superblock is persistent
Update Time : Tue Oct 28 18:08:48 2008
State : clean
Active Devices : 3
Working Devices : 3
Failed Devices : 0
Spare Devices : 0
Chunk Size : 64K
UUID : 0dfcdc4f:61478bb7:ba6711c4:f608aed0
Events : 0.1
Number Major Minor RaidDevice State
0 3 0 0 active sync /dev/hda
1 8 0 1 active sync /dev/sda
2 8 16 2 active sync /dev/sdb
[root@n6re ~]#
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
Additional tests with 3 device raid0 via hdparm.......................
[root@n6re ~]# hdparm -tT /dev/md1
/dev/md1:
Timing cached reads: 2312 MB in 2.00 seconds = 1155.95 MB/sec
Timing buffered disk reads: 288 MB in 3.02 seconds = 95.43 MB/sec
[root@n6re ~]# hdparm -tT /dev/hda
/dev/hda:
Timing cached reads: 2260 MB in 2.00 seconds = 1130.18 MB/sec
Timing buffered disk reads: 94 MB in 3.01 seconds = 31.20 MB/sec
[root@n6re ~]# hdparm -tT /dev/sda
/dev/sda:
Timing cached reads: 2282 MB in 2.00 seconds = 1141.12 MB/sec
Timing buffered disk reads: 130 MB in 3.00 seconds = 43.28 MB/sec
[root@n6re ~]# hdparm -tT /dev/sdb
/dev/sdb:
Timing cached reads: 2228 MB in 2.00 seconds = 1113.98 MB/sec
EDIT: The data confirms the ~93MB/sec read speed of the combo...3 times the slowest, that is three times 31MB/s equals 93mb/s...............
Timing buffered disk reads: 134 MB in 3.01 seconds = 44.55 MB/sec
[root@n6re ~]#
Last edited by lilsirecho (2008-10-29 17:08:00)
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
An item noted in passing..............
The drives used for the raid0 consist of one hda device(CF with adapter) and two other CF cards with sata to CF adapters. Another sata to CF adapter houses the boot device.
The two sata-CF cards in the raid0 config are connected to SATA MASTER ports. The boot sata-CF adapter for boot is connected as a SATA SLAVE device.
There are two master and two slave SATA ports.
Since the boot device loads Faunos into ram and is then inactive, and the raid0 devices are not yet mounted, there is no interference with the raid0 function from the slave device.
Similar performance has been confirmed when running USB boot instead of SATA SLAVE CF device boot .
The hdparm results for the boot SATA SLAVE CF is as follows;
Password:
[root@n6re ~]# hdparm -tT /dev/sdc
/dev/sdc:
Timing cached reads: 2212 MB in 2.00 seconds = 1106.03 MB/sec
Timing buffered disk reads: 130 MB in 3.04 seconds = 42.71 MB/sec
[root@n6re ~]#
A pci-e to SATA adapter might provide better performance since the pci-e bus is 100mhz but would require a second adapter for the SATA to CF conversion. That projects to ~126MB/s for a three CF raid and ~168MB/s for a four device Cf raid0.
I don't have the SATA pci-e adapter..............
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
More info on the raid devices as presently set-up in a raid0 with 3 devices.....................
[root@n6re ~]# hdparm -i /dev/hda
/dev/hda:
Model=, FwRev=20070831, SerialNo=CF CARD 00000603
Config={ HardSect NotMFM Fixed DTR>10Mbs }
RawCHS=15538/16/63, TrkSize=0, SectSize=576, ECCbytes=4
BuffType=DualPort, BuffSize=1kB, MaxMultSect=1, MultSect=off
CurCHS=15538/16/63, CurSects=15662304, LBA=yes, LBAsects=15662304
IORDY=no, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
PIO modes: pio0 pio1 pio2 pio3 pio4
DMA modes: mdma0 mdma1 mdma2
UDMA modes: udma0 udma1 udma2 *udma3 udma4
AdvancedPM=yes: disabled (255)
* signifies the current active mode
[root@n6re ~]# hdparm -i /dev/sda
/dev/sda:
Model=CF 8GB , FwRev=20070831, SerialNo=TSS40037080203114419
Config={ HardSect NotMFM Fixed DTR>10Mbs }
RawCHS=15798/16/63, TrkSize=0, SectSize=576, ECCbytes=4
BuffType=DualPort, BuffSize=1kB, MaxMultSect=1, MultSect=?0?
CurCHS=15798/16/63, CurSects=15924384, LBA=yes, LBAsects=15924384
IORDY=no, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
PIO modes: pio0 pio1 pio2 pio3 pio4
DMA modes: mdma0 mdma1 mdma2
UDMA modes: udma0 udma1 udma2 udma3 udma4 *udma5
AdvancedPM=yes: disabled (255)
Drive conforms to: Unspecified: ATA/ATAPI-4
[root@n6re ~]# hdparm -i /dev/sdb
/dev/sdb:
Model= , FwRev=20070308, SerialNo=CF CARD A10019CC
Config={ HardSect NotMFM Fixed DTR>10Mbs }
RawCHS=15961/16/63, TrkSize=0, SectSize=576, ECCbytes=4
BuffType=DualPort, BuffSize=1kB, MaxMultSect=1, MultSect=?0?
CurCHS=15961/16/63, CurSects=16088688, LBA=yes, LBAsects=16088688
IORDY=no, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
PIO modes: pio0 pio1 pio2 pio3 pio4
DMA modes: mdma0 mdma1 mdma2
UDMA modes: udma0 udma1 udma2 udma3 *udma4
AdvancedPM=yes: disabled (255)
Drive conforms to: Unspecified: ATA/ATAPI-4
* signifies the current active mode
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
Further tests were made on the 3 device combo................
Device hda was interchanged with device sdb...this provided two devices of the same revision number in the SATA DEVICES of sda and sdb.
The result of this change was to provide ......hdparm -t /dev/md1 of 124.35MB/S.....................
The hdparm of hda changed to 43MB/s....both sda and sdb provided 46MB/s.
Hda now shows UDMA4 and both sda and sdb show UDMA5. UDMA4 is capable of supporting 45MB/sec.
I was surprised at the result but can appreciate that revisions do introduce complications. Now, how does one determine the revision before purchase?
This provides almost triple the read speed of a CF card alone(as a true IDE device).
EDIT: Cp of 1GB video file took ~7.8sec. Vlc and depends .pkg.tar.gz loads in 4 secs. Firefox3 and depends loads pkg tar gz in 16secs. These loads from raid0 cachedir as a read from cache write in ram to Desktop.
Last edited by lilsirecho (2008-10-30 01:52:48)
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
Installed x86_64 into intel GA-945GCM/S2C mobo Pentium4 2.8ghz.
Assembled raid0 cache dir as before.
Ran..... hdarm -tT.... on the raid array.
Read speed was reported as 131MB/s.
This may not be accurate in 64 bit. Will test it with a large file..............
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
Created a folder named x86-64 in raid0 array into which I transferred all packages from /var/cache/pacman/pkg. I now have both x86 arch and x86_64 packages cached in raid0.
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
Latest changes include installing two pci-e "mini" adapters for esata. One of the adapters feeds threough an esata to internal sata cable to one raid component of my raid0 sata adapters to CF card. Thus, the system is now running x86-64 with raid0 as the repo(22MB) with all raid cards(3) driven through sata connections.
This change raised the read speed to 137+MB/sec on the raid0 array.
I removed the pci to IDE card and run the system from USB HDD at 28MB/sec read speed.
The second esata "mini" adapter will feed a new flash drive made by ocz throttle esata. The read speeds for these drives is reported as 80MB/s and write speed 30MB/s.
I plan on using the ocz device as a "live Faunos" device for the time being. It should provide speedier start-up and probably faster overlay writes.
I am also hoping to use it in an "all ram" system similar to that being discussed by apaige on another thread.
My "breadboard" system has a low profile...tallest thing is the intel fan for the cpu. The HDD is external USB.
Perhaps by mid-year some Cfast cards will be available which will be an exciting time should the projected speeds be 300MB/S.
Of note, the hdparm readings for the three CF cards in the present setup are all 46MB/S read speed. Since the displayed read speed for the raid0 was 137MB/sec, very little overhead seems apparent. Of course, hdparm is not real world stuff but a general guide.
Previous tests of read speed on large contiguous files has verified the speeds obtained from hdparm. This is not to be considered true for all file sizes.
Some recent discussions of flash drives indicate that fragmentation occurs and speeds decrease when this occurs. Intel has reported methods for reducing this phenom. It is especially true for macro flash cards (multi-layer) but occurs in single layer flash as well. It is introduced by the methods used to control writes.
If one doesn't write often, it would seem fragmentation would be minimized. I think it occurs during reads as well if I recall the details.
Since the methods for controlling the writes is inherently different for manufactured items, YMMV!
Not being an "expert" I only make these comments to bring the subject into view for those interested.
I'll be back!!!
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
Throttle is 8GB and able to run as esata or USB.
Running in Faunos live system as an esata device produces 76MB/sec in hdparm.
This change allows 3 raid0 devices to run from 3 ports, two are pci-e adapter ports and the third is a sataII port.
This allows the remaining sataII port(primary) to be connected to the throttle thru a sataII to esata mating connector cable.
The pci-e esata adapters do not permit the throttle to boot, it can only boot from the sataII port with the connector cable for esata. (JMicron jmb360 pci-e adapters).
This setup provides approximately 4 times the USB 2.0 speed for Faunos flash devices.
I have no more ports to play with unless I "cheat" and use the "Slave" sata ports.
That may be possible for the raid0 devices since they aren't activated in my system except to install from the raid repos and are not mounted until such is required.
Think I will be trying to install all in ram as a next project...from boot to full install uncompressed.
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
Comparng a USB boot with an UDMA CF flash card with the same boot program from SATAII into Faunos.
The boot time with USB was 75 secs.
The boot time with CF card to SATAII adapter was 45 secs from a SATAII port. The hdparm speed is 46MB/sec sequential.
EDIT: The boot time with the ocz throttle in ESATAII was similar to the CF card..43 secs. The hdparm speed for the throttle is 76MB/sec sequential.
Last edited by lilsirecho (2009-02-24 15:32:50)
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
Made a change in the system to allow for pci-e exclusively in the raid0 setup.
Installed an SIIG two-channel internal SATAII card in the pci-e 16x slot. Used jmb360 in the x1 slot.
All perform well in archlinux x86-64 and archlinux x86-64 ext4 and also in Faunos with an earlier kernel.
The drives (3) are recognized as the same SD devices in all boots.
The system Intel945 has two pci-e slots and four possible SataII ports. Using the pci-e slots for the raid0 array permits the mobo SATA port 0 to be used for booting FaunOS via the throttle flash drive. It has a read speed at this time of 75MB/s in FaunOS.
Decided not to pursue the full ram install since FaunOS runs in ram anyhow. My system uses mdadm raid0 setup after booting.
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
The SIIG dual channel pci-e to SATAII permits booting from connected devices.
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline
Regarding Cfast:
Pretec has a 32GB SSD with Cfast config in demo and claims 160MB/Sec with the device. This is similar to speeds with Sata drives.
It is nowhere near the 300MB/s predicted in many circles. It may be affected by overhead problems .
They also plan a 64GB unit to be demoed soon.
No pricing has been found by this reporter.
Prediction...This year will be a very odd year!
Hard work does not kill people but why risk it: Charlie Mccarthy
A man is not complete until he is married..then..he is finished.
When ALL is lost, what can be found? Even bytes get lonely for a little bit! X-ray confirms Iam spineless!
Offline