Home > OS >  Help! Help! Windows2008 x86 every 15 days or so will be stuck
Help! Help! Windows2008 x86 every 15 days or so will be stuck

Time:10-16

Server is dell R720 2 1 t memory 32 gb hard disk group raid 1 CPU E5-2609-1 is mainly run the ERP system, with fewer sql2008r2 use, use at ordinary times, completely no problem performance, sure enough,
A strange question, but often is the machine running about 15 days or so will appear a system without change card, the process of CPU utilization rate is high, ups and downs, volatility is not a fixed process, even if you open my computer, open a folder will card to death, or feel like, like in safe mode, but also some CARDS, restart can't solve, it's no use pulling cable, the temperature is normal, this situation will last about 1 days or so, and will automatically restore, then BuKa, began to think is a virus, with 360, kingsoft has killed, bought a kaspersky behind the enterprise edition, killed a poison, no problem, also consulted a dell engineer, they have seen hardware also said no problem... I also heavy before, reinstall the good for about 3 months later again, ask everybody to help me, help me

CodePudding user response:

Top up · · · · · · · · · · · · · ·

CodePudding user response:

Top up · · · · · · · · · · · · ·

CodePudding user response:

Check the memory footprint, doubt is a memory leak

CodePudding user response:

Overlooked, if restart also such, is not a simple memory leaks,
It is recommended to mount HYPERV server, or KVM, ESXI, pack up the ERP system in virtual machine,
If the problem still, physical machine CARDS, is the problem of physical machine, hardware can be large
If the virtual machine card, that is the problem of ERP,

CodePudding user response:

Friend: good morning,

Ignorance, please listen to not blind those years playing for 286 to i7-920 (now E3-1230 is specifically designed with Windows Storage Server 2012 r2 Workgroup, does not support virtual machine memory is limited to 32 gb; But in double road E5 and 128 gecc memory, PCI - E SSD 730 is too expensive) poor rookie assumed that Syria's way:


Ruled out by the driver generally don't special cause (simple) memory leak, the Windows update trigger. After will be jointly and severally Taskmgr.net process the tide wave open something new for miles when higher CPU usage; Talk about single 32 gb on E5 pick (with E3 level),


Mechanical hard disk cache file fragmentation is the main reason most of the card machine (open drive automatically optimize, slow stochastic instead, speaking, reading and writing); Stand system (C:) on disk "super slow release" third-party lean tools such as cleaning and Dism++, practice patience and all finishing (doing and fastest Win2012R2)

Even two raid 0 1 t plate group, than in previous years but at the P45 Q6600 with Seagate 15000 RPM 300 gb SAS disk under 3.4 ms seek frequent completely finishing (contrast test of patience 7200 RPM single disc 1 t, seek 15 ms), has a cross the experience of jaguar, speed to bother to find plextor M6S - 256 - g the parity of SSD, only feel the road system and Temp, PF on the SSD as fast as memory called zero delay accordingly, continuous and mechanical hard disk read and write level 4 k as web FLASH files occupy the most CPU machine, make the CPU to him from the mouse pointer shake shake until system being dragged to death, SSD greatly shortening the time of hard drives, and even in half CPU usage (even old pentium 4 FLASH advertising amount of web pages)




MHDD according to mechanical hard disk green, red block, the power supply ripple is the killer: (not advertising)

5400 this dish, 7200100, 00 RPM SATA disk bad for two consecutive years (15000 RPM SAS disk and durable), replace magnetic's visit to the great power supply ripple on the high side, on the exquisite ripple set-up DCtoDC chaton G550 compared to listen to, seek lighter and clear, similarly office three same DELL OPTIPLEX - 3020, one of them (supply voltage on the high side, ripple cannot be measured) can make the mechanical hard disk read and write with fatigue and even call delay is obvious, and change to the other two are "gentle",


Since rush 3 entanglements due to mechanical hard disk light from normal read and write flash suddenly to a few seconds long bright blue screen phenomenon, in P45 Q6600 change to be the friend of data recovery can't lend the rookie a rare MHDD sweep green block 7200 RPM single disk 80 g IDE disk all change, contrast is now on a host of big single disk capacity generally have scattered green, red block, especially the green block at the start of the hard disk and tail easily caton crash, even when the red block will collapse to a small percentage of constant velocity, M6S - 128 - g of SSD also slowed to tenths of a bad block, but the second standard format when they show has been reached 400 m/s (standard format in SATA6G constant-velocity)


Sentence: eliminate obstacles, system software to SSD try how cloning system, need not defragment SSD,
  • Related