Hello,
My server (6 months old from burst.net) has been shipped with WDC WD10EARS-00Y hard drive that has a very horrible performance with CentOS5.5 and possible other major Linux bases systems.
Troubleshooting for hours and hours (I am an RHCE myself with +13 experience!) and investigating on forums, there is a consensus that this product doesn't work properly with kernels < 2.6.32. The only solution is to repartition the hard drive using based-8 sectors (i.e, each partition has to begin with sector 64, 128, etc..) while the current partition table is as follow..
I tried to repartition the drive using parted in a way that the first sector is 64 but it failed and insisting on starting with sector 63..Code:Model: ATA WDC WD10EARS-00Y (scsi)
Disk /dev/sda: 1953525167s
Sector size (logical/physical): 512B/512B
Partition Table: msdos
Number Start End Size Type File system Flags
1 63s 401624s 401562s primary ext3
2 401625s 42347339s 41945715s primary ext3
3 42347340s 84293054s 41945715s primary ext3
4 84293055s 1953520064s 1869227010s extended
5 84293118s 117852839s 33559722s logical linux-swap
6 117852903s 126238769s 8385867s logical ext3
7 126238833s 130431734s 4192902s logical ext3
8 130431798s 1953520064s 1823088267s logical ext3
I immediately contacted support for two issues..Code:The closest location we can manage is 63s to 401624s. Is this still acceptable to you?
1- I need the kernel source of the default shipped kernel 2.6.35.4-grsec-hostnoc-4.3.0-x86_64-libata so I can compile VMWare Server against this kernel. The response..
Another response from a different support rep..Code:I apologize but we do not support 3rd party software. Please let me know if you have any further questions or concerns.
Unable to find sources? Are you kidding me?Code:Unfortunately I am unable to find a source download for your specific kernel. If there is anything else needed let me know.
Then they transfered me to customer service department.
2- I have troubleshooted the server for low performance and figured out that the reason is the hard drive (and gave them full explanation, technical data, and online references). The response..
Test what? Did they bother to read a word before wasting ~12 hours of repeating the same tests just to find out eventually that the hard drive doesn't work properly without their default kernel (that they are unable to find it sources) so I can compile VMware against it (since they don't provide support on 3rd party software)?Code:If your hard drive is causing an performance issue we could run a test on the drive for you. However this will take some time considering the drive is large.
The only options I have is to revert back to Centos default kernels, which is assured to have a huge performance downgrades, but then I will be able to run my VMs since I have kernel sources, OR to get their default kernel sources (currently waiting for their reply, transferring me back and forth between departments like support, "build", and customer service!), OR to use a vanilla kernel from kernel.org and to compile everything myself and hope that using a kernel >= 2.6.32 would fix the performance problem in the first place as I reckon.
Seriously, if this is the way it goes with experienced customers how do things go for the average joe?
Why in the first place I am delivered a server with a hard drive that is known to have issues? In the end of the day if using a kernel that is above 2.6.31 failed to resolve the issue I am having no option but to ask for a hard drive imaging and replacement but question is will they do it for free when they can't acknowledge that there is an issue with their delivered configurations?
I had problems in the past with burst.net on customer support level, sales level, billing level, you name it! But I was patient enough with them to solve all the issue (in about a month) and now yet another problem with support?
The only reason for me to keep the server and not running for any other competitor that I already use is that my business requires for my clients sites IP addresses to be the same. Plus they charged me like a month and a half in advance and my account balance there is like post-paid hundreds of dollars (that's another story and was the only solution to get rid of the sales/billings issues they have).
Am I being harsh or not logic with them? Too much expectations perhaps?
I have been in the hosting industry for years and dealt with almost every hosting company like iweb, ev1 before becoming TP then orbit then now SL, and even european based hosts like easyspeedy, you name it! I never seen such problems with a company on ALL levels.
It's just that solving all the problems with patience on other levels than TECHNICAL SUPPORT gives me hope that the latter will know how to do their job (although there is no patience one can afford when you have a technical issue). Or maybe I am mistaken and have to find a way to run from this host?
Respectfully,
Hossam Hossny