New answers tagged

1

KTD-WS667LPQ is 'Quad' Rank FBDIMMs unsupported by HP ProLiant DL380 G5 Only Single Rank and Dual Rank modules supported. KTM5780 and KTH-XW667 are Dual Rank FBDIMMs, so it will work. For complete manual, click here.


1

There are several guidelines that you need to adhere specifically for your server. The population method is based on Ranks, assuming you have single and dual ranked modules, and 16 GB Quad Ranks, the following will apply You must install the 16 GB modules in the white levered slots: 1) 16 GB's Bank #1 (A1, A2, A3, B1, B2, B3) 2) 4 GB's Bank#2 (A4, A5, ...


1

I want to leave this note as a comment, but I think my concern has been sufficiently (like 95%) alleviated, so I'm answering myself. the comments to this answer show some anecdotes where people find a very low rate of data errors. A rate that is certainly low enough for me to tolerate, as I do not require some kind of large-number-of-nines uptime guarantee. ...


0

Try use the Compressed References parameters to ensure that you are taking full advantage of your 32 GB of memory. Depending on your version of JRockit, you might be using 4 GB compressed references thus exhausting the 4 GB memory space that is shared with other processes. This article has a good discussion of using this parameter: ...


0

Leave the swap at 6GB. If it worked before with 8GB+6GB it will also work with 16GB+6GB. More info here: http://askubuntu.com/questions/62073/how-to-decide-on-swap-size


3

The solution was to Upgrade to kernel-2.6.32-573.7.1.el6 or higher. A simple yum update and reboot should be all it takes. Below is a quote of the bug report of kernel-2.6.32-573.1.1.el6.x86_64 that caused the Swap free to be greater than the swap total: A previous change in the get_swap_page() locking removed the use of the swap_lock spinlock. This ...


0

I don't know if you can prevent caching, but you could regularly ask linux to drop the cache. The way to do it is explained there: http://unix.stackexchange.com/questions/87908/how-do-you-empty-the-buffers-and-cache-on-a-linux-system If you sync just before the echo to /proc/sys/vm/drop_caches command, it should free a lot of ram. Linux will not actively ...


0

I found this explanation from Mugurel Sumanariu very clear: VIRT stands for the virtual size of a process, which is the sum of memory it is actually using, memory it has mapped into itself (for instance the video card’s RAM for the X server), files on disk that have been mapped into it (most notably shared libraries), and memory shared with other ...


2

First of all, these httpd children don't really use the amount of memory that top or ps reports in. Unix uses a complicated VM copy-on-write model, so when a process forks both parent and a child think they use the same amount of memory, so does top and ps. But in fact they share (please don't mix it with IPC shared) memory until some of them modifies a VM ...


0

Thanks to this question, I decided to get the BIOS on the machines updated, and things seem to now be working as expected. Quite probably due to different ranking/banking layout on the DIMMs, which the BIOS update introduced added support for. EDIT: Would also like to thank @ewwhite for his input on the official HP part numbers. Money not being an issue, ...


1

This shows that both NUMA nodes allocated 13.1 GB of memory of 16GB. OS should take some memory for itself but there is to big 3GB of memory difference. Not quite. It shows 8003Mb available on Node 0 and 8192Mb available on Node 1, which is different to "allocated". Also to note: 6479/8003 = 0.80957 6632/8192 = 0.80957 Coincidence? I think not. It ...


3

If you want to populate RAM for a server like this, I can't see any barriers to you doing the right thing and using the RAM specifically meant for this model; specifically HP part # 397415-B21, which is comprised of 2 x 4GB modules. The DIMMs are only a few dollars for this very old server model.


0

How are you populating your slots? From what i gather on the G5 the memory should be skipping a slot for matching dimms. EX. KTD-WS667LPQ in slots 1 and 3, KTM5780 in slot 2 and KTH-XW667 in slot 4, Samsung M395T5163QZ4 in slots 5 and 7 Try each of those sets by themselves to see if they work fine, if they do work fine all by themselves then try adding ...


76

Obviously, memtest86+ cannot test the memory region which currently contains the memtest86+ executable code (but if there are memory errors in that region, it is very likely that the test itself will crash). However, memtest86+ is able to relocate its own code to a different address at runtime, and by using this trick it is able to test all memory which it ...


5

No, memtest can't test its own memory. However, it is so small (only some KB) that it hardly matters. EDIT: this statement is wrong since, as stated in the selected answer, memtest can dynamically relocate itself to test all user addressable memory. -- In theory, modern processor can, at boot time, configure part of their cache as programmable memory, from ...


1

By default 64MB is allocated. You can increase it using the -m option. For example, to increase it to 128MB add it in the memcached.conf as below in a new line and restart memcached: -m 128 Check: Using memcached and Memcached Installation and Configuration with PHP on Debian server


0

The timings with DDR sdram are a multiple from 8 AND from 2. From what I remember, DDR means Double Data Rate (please check this, because I could not find the information to check again). That memory is able to transfer both ways (in & out) in the mean time. The reason is easy. The DDR is timed twice faster than the given bus frequency. Please do not ...


1

I never did find a clear interpretation of the mcelog data, but my best guess worked out, and I figured I should follow up for posterity. I assumed CPU 1 meant the second CPU, helpfully labeled as 2 on the motherboard diagram. I assumed MEMORY CONTROLLER MS_CHANNEL3_ERR indicated channel 3 on that CPU's memory controller. As above, that channel controls ...


3

The R410 is a old server. Size must match in the bank. (tri-channel and dual bank), (but slot 4 is a single slot bank) Bank A1 & A2 = 2 x 16G Bank B1 & B2 = 2 x 8G or Slot A4 = 1 x 16G Slot B4 = 1 x 16G You will got a "Unsupported memory configuration. DIMM mismatch across slots detected" if the memory does not fit across slot 1 to 3. If you use ...


2

The situation still seems quite unclear so I don't think your questions can be answered directly, but here is some relatively recent information as a partial answer. For news, follow the rowhammer-discuss mailing list. I'm not sure it is possible at present with public information to avoid buying vulnerable RAM, nor to easily predict failure rates in ...


0

I think there may be some confusion just based on the title of the question. If you just mean the average desktop PC, then that is usually based on a platform that doesn't even have ECC support. If you mean a workstation class computer, then it quite likely comes with ECC memory whether you care about it or not. Overall, the workstation class is typically ...


2

This looks like memory ballooning is active. This occurs when the hypervisor is under memory pressure; it claims unused memory from some guests to allocate to others. Generally the memory is automatically returned when the guest asks for it back. This is commonly seen with VPS providers who overcommit their resources, knowing that most users will not use ...


0

It's not a super satisfying answer, but enabling memory and swap accounting in the kernel seems to have reined in the excessive memory usage.



Top 50 recent answers are included