|
Home > Archive > Data Storage > September 2004 > What's reasonable RAID 5 performance?
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
What's reasonable RAID 5 performance?
|
|
| kenw@kmsi.net 2004-09-06, 5:45 pm |
| I have a nice new server with 15,000RPM SCSI drives in a hardware RAID 5
configuration. I may be wrong, but I don't think it's performing properly
at all. The question is, what's reasonable?
When using large (>500MB) files to swamp out cache effects, I'm getting
roughly 12MB/sec (it varies quite a bit) write performance and maybe 200
MB/sec, when measured with IOzone.
Measuring with batch 'copy' commands, I'm getting 40MB/sec read-only (copy
to NUL and about 14MB/sec copying back to the same array. One of the
challenges has been getting consistent results; not sure why.
These numbers strike me as being OTL (out to lunch) for such
high-performance drives and array controller.
The array consists of five Seagate Cheetah ST336753LC 15,000RPM drives
connected via a 320MB/sec SCSI interface. The controller is an Intel
SRCU42X as provided by Intel, i.e., the standard 128MB cache, and no
on-board battery (the server has a UPS and redundant power supplies
connected to separate power sources) which means the controller will do
write-through, but not write-back, caching. No override is available.
Intel claims that this is the optimum configration for the controller, and
that more RAM or the battery pack will not help performance significantly.
The motherboard, BTW, is an Intel SE7501HG2 with dual 2.8GHz Xeons and 2GB
of RAM.
There must be thousands of RAID 5 arrays out there very similar to this
one. _Somebody_ must know. Are these performance figures reasonable, or
not?
Ken Wallewein
K&M Systems Integration
Phone (403)274-7848
Fax (403)275-4535
kenw@kmsi.net
www.kmsi.net
| |
| Eric Gisin 2004-09-06, 5:45 pm |
| <kenw@kmsi.net> wrote in message
news:re4pj0hj19blha05h2283n93jf7civi3kv@
4ax.com...
> I have a nice new server with 15,000RPM SCSI drives in a hardware RAID 5
> configuration. I may be wrong, but I don't think it's performing properly
> at all. The question is, what's reasonable?
>
> When using large (>500MB) files to swamp out cache effects, I'm getting
> roughly 12MB/sec (it varies quite a bit) write performance and maybe 200
> MB/sec, when measured with IOzone.
>
> Measuring with batch 'copy' commands, I'm getting 40MB/sec read-only (copy
> to NUL and about 14MB/sec copying back to the same array. One of the
> challenges has been getting consistent results; not sure why.
>
The destination file has to be contiguous to get proper results. This is not
likely with cmd's copy or Sandra. Xcopy is if it can preallocate contig free
space, but it will be seek bound if you have a single array.
Create a 10MB temp file (so it stays in cache), and do "copy/b big+big+(18
more) bigger". Or xcopy from a server if you have GB ethernet.
| |
| Arno Wagner 2004-09-06, 5:45 pm |
| In comp.sys.ibm.pc.hardware.storage kenw@kmsi.net wrote:
> I have a nice new server with 15,000RPM SCSI drives in a hardware RAID 5
> configuration. I may be wrong, but I don't think it's performing properly
> at all. The question is, what's reasonable?
> When using large (>500MB) files to swamp out cache effects, I'm getting
> roughly 12MB/sec (it varies quite a bit) write performance and maybe 200
> MB/sec, when measured with IOzone.
One data-point: On a software RAID5 with 2.6.7 and 5 * Maxtor
200GB DiamondMax 9 plus I get 22MB/s large file write performance
(measured with 1GB data file) and 65MB/s read performance. That is with
ext3 journalling file system, which also journals data, not only
metadata.
> Measuring with batch 'copy' commands, I'm getting 40MB/sec read-only (copy
> to NUL and about 14MB/sec copying back to the same array. One of the
> challenges has been getting consistent results; not sure why.
> These numbers strike me as being OTL (out to lunch) for such
> high-performance drives and array controller.
I would say the performance is rather embarassing when a software
RAID on half as fast disks performs massively better. However I
recenly made the mistake of buying an adaptec SATA RAID controller.
Also slower than software RAID. I also recently talked to some
guy running huge usenet servers: They also have noted that hardware
RAID is now slower than software RAID. As soon as Linux supports
ATA/SATA hotplugging the last advantage of hardware RAID will
be gone.
Arno
--
For email address: lastname AT tik DOT ee DOT ethz DOT ch
GnuPG: ID:1E25338F FP:0C30 5782 9D93 F785 E79C 0296 797F 6B50 1E25 338F
"The more corrupt the state, the more numerous the laws" - Tacitus
| |
| Ron Reaugh 2004-09-07, 2:46 am |
|
"Eric Gisin" <ericgisin@graffiti.net> wrote in message
> Create a 10MB temp file (so it stays in cache), and do "copy/b big+big+(18
> more) bigger". Or xcopy from a server if you have GB ethernet.
Gigabit isn't fast enough here.
| |
| Ron Reaugh 2004-09-07, 2:46 am |
|
"Arno Wagner" <me@privacy.net> wrote in message news:2q47hiFqn6sfU1@uni-
> Also slower than software RAID. I also recently talked to some
> guy running huge usenet servers: They also have noted that hardware
> RAID is now slower than software RAID.
Some HW RAID may be slower but not the right stuff configured properly. SW
RAID is moving in on most the territories though.
| |
| Rob Turk 2004-09-07, 2:48 am |
| <kenw@kmsi.net> wrote in message
news:re4pj0hj19blha05h2283n93jf7civi3kv@
4ax.com...
[SNIP]
>The controller is an Intel
> SRCU42X as provided by Intel, i.e., the standard 128MB cache, and no
> on-board battery (the server has a UPS and redundant power supplies
> connected to separate power sources) which means the controller will do
> write-through, but not write-back, caching. No override is available.
>
If your 12MB/s is really what you get, then flushing your 128MB cache takes
10 seconds. If some idiot decides to push the power button on the server, it
will switch off before your cache is flushed. Maybe not such a bad idea to
get the battery option anyway??
Rob
| |
| kenw@kmsi.net 2004-09-07, 7:45 am |
| "Ron Reaugh" <rondashreaugh@att.net> wrote:
>"Eric Gisin" <ericgisin@graffiti.net> wrote in message
>
>Gigabit isn't fast enough here.
Sure it is. It's far faster than the throughput I'm getting from the array
right now, and faster than a 32-bit PCI bus (the RAID server's 64). As it
happens, the only system I currently have to trade files with has a 32bit
PCI, and when I watch network utilization, the bottleneck is obvious.
A gigabit network should be able to approach 100MB/sec -- say, at least 80.
If I was getting that from my RAID array, I'd be happy.
BTW, the copy append idea in Eric's message is a great idea. It
effectively lets me do write-only write performance testing, almost the
reverse of my copy-to-NUL read test. Cool!
Unfortunately, none of this either confirms or denies whether my current
RAID 5 performance is reasonable.
/kenw
Ken Wallewein
K&M Systems Integration
Phone (403)274-7848
Fax (403)275-4535
kenw@kmsi.net
www.kmsi.net
| |
| Ron Reaugh 2004-09-07, 5:45 pm |
|
<kenw@kmsi.net> wrote in message
news:8ubrj0l2h2d2r32a7ueg18cfsajcb7gfg5@
4ax.com...
> "Ron Reaugh" <rondashreaugh@att.net> wrote:
>
big+big+(18[vbcol=seagreen]
>
> Sure it is. It's far faster than the throughput I'm getting from the
array
> right now,
You just contradicted yourself overall. Your stated goal is "should be
getting". For that gigabit is NOT fast enough. What about the 200?
> and faster than a 32-bit PCI bus (the RAID server's 64).
No, gigabit is about the same speed at peak of 32 bit 33 Mhz PCI.
> As it
> happens, the only system I currently have to trade files with has a 32bit
> PCI, and when I watch network utilization, the bottleneck is obvious.
Rethink what you are watching.
> A gigabit network should be able to approach 100MB/sec -- say, at least
80.
That's what I've said and 32 bit 33.3 Mhz PCI does 133.3 MB/sec.
> If I was getting that from my RAID array, I'd be happy.
What about the 200?
> BTW, the copy append idea in Eric's message is a great idea. It
> effectively lets me do write-only write performance testing, almost the
> reverse of my copy-to-NUL read test. Cool!
>
> Unfortunately, none of this either confirms or denies whether my current
> RAID 5 performance is reasonable.
| |
| Folkert Rienstra 2004-09-07, 5:45 pm |
|
"Rob Turk" <_wipe_me_r.turk@chello.nl> wrote in message news:8Ub%c.65750$C7.51368@amsnews05.chello.com
> <kenw@kmsi.net> wrote in message news:re4pj0hj19blha05h2283n93jf7civi3kv@
4ax.com...
> [SNIP]
>
> If your 12MB/s is really what you get, then flushing your 128MB cache
> takes 10 seconds.
Oh?
What about the "the controller will do write-through, but not write-back, caching"?
> If some idiot decides to push the power button on the server, it
> will switch off before your cache is flushed.
> Maybe not such a bad idea to get the battery option anyway??
>
> Rob
| |
| Ron Reaugh 2004-09-07, 5:45 pm |
|
"Folkert Rienstra" <see_reply-to@myweb.nl> wrote in message
news:2q6hdjFredu7U1@uni-berlin.de...
>
> "Rob Turk" <_wipe_me_r.turk@chello.nl> wrote in message
news:8Ub%c.65750$C7.51368@amsnews05.chello.com
news:re4pj0hj19blha05h2283n93jf7civi3kv@
4ax.com...[vbcol=seagreen]
do[vbcol=seagreen]
>
> Oh?
> What about the "the controller will do write-through, but not write-back,
caching"?
Read the thread before blathering. The controller WILL do write-back when
it has a battery.
[vbcol=seagreen]
| |
| Farkash 2004-09-09, 5:25 am |
| http://www.outsupport.org/forum/viewtopic/985.html
http://forum.shrc.org/english/bb/Fo...TML/000030.html
http://www.fpa.org/thread_msg2597/t...TML/000194.html
***Special CONFIRMED Report. ****Assassins; who put Al-Qaeda to Shame.
The Number three most powerful man , after Dick Cheney & G.W. BUSH .
ALL Ariel Sharon's servants , Thugs & Murderers.
Karl ROVE & Ariel Sharon banking on their Syrian killers & Murderers &
Special Syrian Assassins of Assef Shawkat & Roustom Ghazali Working for
Sharon and the NEOCONS.
Special ICC Investigation; The Hague NL. & Belgium .
****************************************
http://www.onlinejournal.com/Specia...2104madsen.html
Karl Rove's White House " Murder, Inc."
By Wayne Madsen .
Online Journal Contributing Writer .
SEP, 2004- On September 15, 2001, just four days after the 9-11 attacks,
CIA Director George Tenet provided President [sic] Bush with a Top Secret
"Worldwide Attack Matrix"-a virtual license to kill targets deemed to be a
threat to the United States in some 80 countries around the world. The Tenet
plan, which was subsequently approved by Bush, essentially reversed the
executive orders of four previous U.S. administrations that expressly
prohibited political assassinations.
According to high level European intelligence officials, Bush's counselor,
Karl Rove, used the new presidential authority to silence a popular Lebanese
Christian politician who was planning to offer irrefutable evidence that
Israeli Prime Minister Ariel Sharon authorized the massacre of hundreds of
Palestinian men, women, and children in the Beirut refugee camps of Sabra
and Shatilla in 1982. In addition, Sharon provided the Lebanese forces who
carried out the grisly task. At the time of the massacres, Elie Hobeika was
intelligence chief of Lebanese Christian forces in Lebanon who were battling
Palestinians and other Muslim groups in a bloody civil war. He was also the
chief liaison to Israeli Defense Force (IDF) personnel in Lebanon. An
official Israeli inquiry into the massacre at the camps, the Kahan
Commission, merely found Sharon "indirectly" responsible for the slaughter
and fingered Hobeika as the chief instigator.
The Kahan Commission never called on Hobeika to offer testimony in his
defense. However, in response to charges brought against Sharon before a
special war crimes court in Belgium, Hobeika was urged to testify against
Sharon, according to well-informed Lebanese sources. Hobeika was prepared to
offer a different version of events than what was contained in the Kahan
report. A 1993 Belgian law permitting human rights prosecutions was unusual
in that non-Belgians could be tried for violations against other
non-Belgians in a Belgian court. Under pressure from the Bush
administration, the law was severely amended and the extra territoriality
provisions were curtailed.
Hobeika headed the Lebanese forces intelligence agency since the mid- 1970s
and he soon developed close ties to the CIA. He was a frequent visitor to
the CIA's headquarters at Langley, Virginia. After the Syrian invasion of
Lebanon in 1990, Hobeika held a number of cabinet positions in the Lebanese
government, a proxy for the Syrian occupation authorities. He also served in
the parliament. In July 2001, Hobeika called a press conference and
announced he was prepared to testify against Sharon in Belgium and revealed
that he had evidence of what actually occurred in Sabra and Shatilla.
Hobeika also indicated that Israel had flown members of the South Lebanon
Army (SLA) into Beirut International Airport in an Israeli Air Force C130
transport plane. In full view of dozens of witnesses, including members of
the Lebanese army and others, SLA troops under the command of Major Saad
Haddad were slipped into the camps to commit the massacres. The SLA troops
were under the direct command of Ariel Sharon and an Israeli Mossad agent
provocateur named Rafi Eitan. Hobeika offered evidence that a former U.S.
ambassador to Lebanon was aware of the Israeli plot. In addition, the IDF
had placed a camera in a strategic position to film the Sabra and Shatilla
massacres. Hobeika was going to ask that the footage be released as part of
the investigation of Sharon.
After announcing he was willing to testify against Sharon, Hobeika became
fearful for his safety and began moves to leave Lebanon. Hobeika was not
aware that his threats to testify against Sharon had triggered a series of
fateful events that reached well into the White House and Sharon's office.
On January 24, 2002, Hobeika's car was blown up by a remote controlled bomb
placed in a parked Mercedes along a street in the Hazmieh section of Beirut.
The bomb exploded when Hobeika and his three associates, Fares Souweidan,
Mitri Ajram, and Waleed Zein, were driving their Range Rover past the
TNT-laden Mercedes at 9:40 am Beirut time. The Range Rover's four passengers
were killed in the explosion. In case Hobeika's car had taken another route
through the neighborhood, two additional parked cars, located at two other
choke points, were also rigged with TNT. The powerful bomb wounded a number
of other people on the street. Other parked cars were destroyed and
buildings and homes were damaged. The Lebanese president, prime minister,
and interior minister all claimed that Israeli agents were behind the
attack.
It is noteworthy that the State Department's list of global terrorist
incidents for 2002 worldwide failed to list the car bombing attack on
Hobeika and his party. The White House wanted to ensure the attack was
censored from the report. The reason was simple: the attack ultimately had
Washington's fingerprints on it.
High level European intelligence sources now report that Karl Rove
personally coordinated Hobeika's assassination. The hit on Hobeika employed
Syrian intelligence agents. Syrian President Bashar Assad was trying to
curry favor with the Bush administration in the aftermath of 9-11 and was
more than willing to help the White House. In addition, Assad's father,
Hafez Assad, had been an ally of Bush's father during Desert Storm, a period
that saw Washington give a "wink and a nod" to Syria's occupation of
Lebanon. Rove wanted to help Sharon avoid any political embarrassment from
an in absentia trial in Brussels where Hobeika would be a star witness. Rove
and Sharon agreed on the plan to use Syrian Military Intelligence agents to
assassinate Hobeika. Rove saw Sharon as an indispensable ally of Bush in
ensuring the loyalty of the Christian evangelical and Jewish voting blocs in
the United States. Sharon saw the plan to have the United States coordinate
the hit as a way to mask all connections to Jerusalem.
The Syrian hit team was ordered by Assef Shawkat, the number two man in
Syrian military intelligence and a good friend and brother in law of Syrian
President Bashar Assad. Assad's intelligence services had already cooperated
with U.S. intelligence in resorting to unconventional methods to extract
information from al Qaeda detainees deported to Syria from the United States
and other countries in the wake of 9-11. The order to take out Hobeika was
transmitted by Shawkat to Roustom Ghazali, the head of Syrian military
intelligence in Beirut. Ghazali arranged for the three remote controlled
cars to be parked along Hobeika's route in Hazmieh; only few hundred yards
from the Barracks of Syrian Special Forces which are stationed in the area
near the Presidential palace , the ministry of Defense and various
Government and officers quarters . This particular area is covered 24/7 by a
very sophisticated USA multi-agency surveillance system to monitor Syrian
and Lebanese security activities and is a " Choice " area to live in for its
perceived high security .... [Courtesy of the Special Collections Service.]
... SCS... ; CIA & NSA & DIA....
The plan to kill Hobeika had all the necessary caveats and built-in denial
mechanisms. If the Syrians were discovered beforehand or afterwards, Karl
Rove and his associates in the Pentagon's Office of Special Plans would be
ensured plausible deniability.
Hobeika's CIA intermediary in Beirut, a man only referred to as "Jason" by
Hobeika, was a frequent companion of the Lebanese politician during official
and off-duty hours. During Hobeika's election campaigns for his
parliamentary seat, Jason was often in Hobeika's office offering support and
advice. After Hobeika's assassination, Jason became despondent over the
death of his colleague. Eventually, Jason disappeared abruptly from Lebanon
and reportedly later emerged in Pakistan.
Karl Rove's involvement in the assassination of Hobeika may not have been
the last "hit" he ordered to help out Sharon. In March 2002, a few months
after Hobeika's assassination, another Lebanese Christian with knowledge of
Sharon's involvement in the Sabra and Shatilla massacres was gunned down
along with his wife in Sao Paulo, Brazil. A bullet fired at Michael Nassar's
car flattened one of his tires. Nassar pulled into a gasoline station for
repairs. A professional assassin, firing a gun with a silencer, shot Nassar
and his wife in the head, killing them both instantly. The assailant fled
and was never captured. Nassar was also involved with the Phalange militia
at Sabra and Shatilla. Nassar was also reportedly willing to testify against
Sharon in Belgium and, as a nephew of SLA Commander General Antoine Lahd,
may have had important evidence to bolster Hobeika's charge that Sharon
ordered SLA forces into the camps to wipe out the Palestinians.
Based on what European intelligence claims is concrete intelligence on
Rove's involvement in the assassination of Hobeika, the Bush administration
can now add political assassination to its laundry list of other misdeeds,
from lying about the reasons to go to war to the torture tactics in
violation of the Geneva Conventions that have been employed by the Pentagon
and "third country" nationals at prisons in Iraq and Guantanamo Bay.
Wayne Madsen is a Washington, DC-based investigative journalist and
columnist. He served in the National Security Agency (NSA) during the Reagan
administration and wrote the introduction to Forbidden Truth. He is the
co-author, with John Stanton, of "America's Nightmare: The Presidency of
George Bush II." His forthcoming book is titled: "Jaded Tasks: Big Oil,
Black Ops, and Brass Plates." Madsen can be reached at:
WMadsen777@aol.com
This is some of the evidence for you and for the World ....
~~~encrypted/logs/access ~~~
Not to mention hundreds of private companies and governments. Anyway...
Lines 10-36
of my logfiles show a lot of interest in this article:
# grep sid=1052 /encrypted/logs/access_log|awk '{print $1,$7}'|sed -n
'10,36p'
spb-213-33-248-190.sovintel.ru /modules.php?name=News&file=article&sid=1052
ext1.shape.nato.int /modules.php?name=News&file=article&sid=1052
server1.namsa.nato.int /modules.php?name=News&file=article&sid=1052
ns1.saclantc.nato.int /modules.php?name=News&file=article&sid=1052
bxlproxyb.europarl.eu.int /modules.php?name=News&file=article&sid=1052
wdcsun18.usdoj.gov /modules.php?name=News&file=article&sid=1052
wdcsun21.usdoj.gov /modules.php?name=News&file=article&sid=1052
tcs-gateway11.treas.gov /modules.php?name=News&file=article&sid=1052
tcs-gateway13.treas.gov /modules.php?name=News&file=article&sid=1052
relay1.ucia.gov /modules.php?name=News&file=article&sid=1052
relay2.cia.gov /modules.php?name=News&file=article&sid=1052
relay2.ucia.gov /modules.php?name=News&file=article&sid=1052
n021.dhs.gov /modules.php?name=News&file=article&sid=1052
legion.dera.gov.uk /modules.php?name=News&file=article&sid=1052
gateway-fincen.uscg.mil /modules.php?name=News&file=article&sid=1052
crawler2.googlebot.com /modules.php?name=News&file=article&sid=1052
crawler1.googlebot.com /modules.php?name=News&file=article&sid=1052
gateway101.gsi.gov.uk /modules.php?name=News&file=article&sid=1052
gate11-quantico.nmci.usmc.mil /modules.php?name=News&file=article&sid=1052
gate13-quantico.nmci.usmc.mil /modules.php?name=News&file=article&sid=1052
fw1-a.osis.gov /modules.php?name=News&file=article&sid=1052
crawler13.googlebot.com /modules.php?name=News&file=article&sid=1052
fw1-b.osis.gov /modules.php?name=News&file=article&sid=1052
bouncer.nics.gov.uk /modules.php?name=News&file=article&sid=1052
beluha.ssu.gov.ua /modules.php?name=News&file=article&sid=1052
zukprxpro02.zreo.compaq.com
/modules.php?name=News&file=article&sid=1052....
To be continued ....
HOLLYWOOD FL.... ATTA & Aris2FatCat
DENVER CO
ART STUDENTS...
MOOVERS INC.@IL
Lakam & Mr.X. MEGA
Etc. Etc. |
|
|
|
|