PLE is too high (larger than uptime...)

Question

PLE is too high (larger than uptime...)

az-sqlsc

SSC Enthusiast

Points: 175
More actions
September 1, 2016 at 3:51 am

#321036

Hi,
Server has been restart this morning, i got
[font="System"]object_nameSQLServer:Buffer Manager
counter_namePage life expectancy
UPTIME_MIN312
PLE_SECS3600143
PLE_MINS60002[/font]
with this request
SELECT [object_name]
,[counter_name]
, UPTIME_MIN = CASE WHEN[counter_name]= 'Page life expectancy'
THEN (SELECT DATEDIFF(MI, MAX(login_time),GETDATE())
FROM master.sys.sysprocesses
WHERE cmd='LAZY WRITER')
ELSE ''
END
, [cntr_value] AS PLE_SECS
,[cntr_value]/ 60 AS PLE_MINS
FROM sys.dm_os_performance_counters
WHERE [object_name] LIKE '%Manager%'
AND[counter_name] = 'Page life expectancy'
and same PLE value with BrentOzar and Glenn Berry scripts
Did i misunderstood something ? google didn't help me
sql version is 10.50.6529.0, on vmware with 4 vcpu and 16 GB ram
thx
Az'

Viewing 7 posts - 1 through 6 (of 6 total)

You must be logged in to reply to this topic. Login to reply

spaghettidba SSC Guru Points: 105732 More actions · Answer 1

PLE is a counter that represents the number of seconds a data page is expected to live in the buffer cache before it is evicted to make room for other pages.

This is everything you need to know about PLE: it doesn't matter whether it's higher than your uptime, there is no correlation between the two.

-- Gianluca Sartori

Lynn Pettis SSC Guru Points: 442467 More actions · Answer 2

spaghettidba (9/1/2016)
PLE is a counter that represents the number of seconds a data page is expected to live in the buffer cache before it is evicted to make room for other pages.
This is everything you need to know about PLE: it doesn't matter whether it's higher than your uptime, there is no correlation between the two.

Not to go off-topic, but it is nice to see you around after the earthquake in Italy. Glad to know you are okay.

spaghettidba SSC Guru Points: 105732 More actions · Answer 3

Lynn Pettis (9/1/2016)
spaghettidba (9/1/2016)
PLE is a counter that represents the number of seconds a data page is expected to live in the buffer cache before it is evicted to make room for other pages.
This is everything you need to know about PLE: it doesn't matter whether it's higher than your uptime, there is no correlation between the two.
Not to go off-topic, but it is nice to see you around after the earthquake in Italy. Glad to know you are okay.

Thanks Lynn. I live hundreds of Km away from the earthquake.

OP: Sorry for hijacking your question... 🙂

-- Gianluca Sartori

az-sqlsc SSC Enthusiast Points: 175 More actions · Answer 4

OP: Sorry for hijacking your question... 🙂

That's all right 🙂

So i assume it's a first guess ? the sql loads X GB in Buffer in Y minutes on Z GB available, it makes an expectancy guess and now increments this ple normally until pressure?

new counter are

UPTIME_MINPLE_SECSPLE_MINS

641 361989560331

before

312 60002

so exactly 59690 + Uptime for the PLE, makes me suspicious

Az'

spaghettidba SSC Guru Points: 105732 More actions · Answer 5

I have no idea how SQL Server calculates that number, especially when no page has ever been flushed from the cache: I can only tell you what it means.

I don't think you should try to correlate it with SQL Server uptime.

-- Gianluca Sartori

az-sqlsc SSC Enthusiast Points: 175 More actions · Answer 6

ok thank you for your help,

the keyword is 'expectancy' so i assume high value doesn't mean anything more than in good condition.