Errors and problems of the Big Three servers: part one. Dell
Being engaged in the restored (refurbished) servers , we have gained a decent experience in dealing with problems of different brands, which we want to share with the example of the Big Three machines: Dell, HP and IBM. To no one was hurt, we will go in alphabetical order.
In this post we collected feedback from our engineers who have extensive experience in diagnosing, repairing and maintaining servers - each described error was solved by their sweat and blood. This information is not useful if you are serviced under warranty, but may nevertheless be useful.
The problem when updating software is relevant for equipment DELL older than 3-5 years. Recall that all manipulations with the Dell server are made using Lifecycle, which is built into the server software, as well as the iDRAC.
')
Here, in the standard scenario, there are two options. First, the update is in manual mode, from a flash card - in case there is no Internet connected to the server. Then go to the Lifecycle and update from a flash drive. If something does not work, we connect the server to the Internet and update, or we go through the iDRAC and restore the initial settings, install the necessary updates.
However, if the server is old enough, the problems are waiting almost immediately: when trying to install software over the Internet from the Lifecycle update and diagnostics system. The reason for this problem is that the manufacturer initially, at the dawn of this system, did not use the security protocol on the FTP server to update the firmware of the components of its servers. Now, when such certificates of protection are introduced, the old equipment of three to five years ago is sometimes difficult to update via the Internet.
The problem is solved quite simply and many of you have probably already guessed how this can be done: update the software, namely Lifecycle, but first update the IDRAC system. Updates are available for free on the official Dell technical support site. ( Check warranty and download updates to your Dell server ).
After this manipulation, you can conveniently download all the drivers and other updates via the Internet, regardless of the year of release of your server - if, of course, there is a set of updates for it on the manufacturer's website.
Recall that the iDRAC is presented in the options "standard", "express" and "enterprise". The standard comes with the purchase of the server. If we order a complete set of "express", then on the motherboard of the server we are put the appropriate licensed chip. Thus, we get a new port that allows us to receive more informative data via the Web.
With a standard license, when we connect to the iDRAC system on a work computer via a browser, we can see an overview of the system information on a standard tab, server service tag, model, and nothing more.
Via express, i.e. if we have version 7 of iDRAC, we can already update the BIOS, Lifecycle, etc. directly through the iDRAC remote access system. In earlier versions of the system, unfortunately, this feature is not provided: it can only be updated via a flash drive.
Enterprise is the most complete set of server management functions available through the iDRAC port, which allows you to fully monitor updates, a full list of components, and the current state of the server. This is a convenient and functional kit, but we managed to find in it one small but significant drawback.
After updating the version of iDRAC Enterprise 1.30.30 or 1.45.45, the system stops initializing and does not start due to incorrect settings. As a result, if we enter the iDRAC in the settings via the BIOS, we can neither turn it off, nor do anything to restore it or reboot it.
For most users, the only way out in this situation is to contact technical support. If your server is under warranty, then specialists will come and replace the motherboard. We, in turn, faced this problem and, oddly enough, a completely standard solution, described on the official Dell forum, helped: the server had to be de-energized for 5 minutes, pull out the power supplies and disconnect all the wires. True, if this does not help, then you still have to replace the motherboard.
Error E1410 is issued after the upgrade: you have added another processor, but for some reason the server does not start with it. Rather, it turns on, the indicators light up alternately, it makes a loud noise, but there is nothing on the screen and it writes an error on the display of the diagnostic panel.
The problem in this case is that on the socket CPU 1, the socket legs are bent, which are adjacent to the processor - either one or several at once. In the image you can see the working legs of the socket (on the left) and the place where they are bent (on the right). The legs can be manually straightened, the problem will be settled.
As you have already guessed, with this error, the second processor does not start because some of the functionality of the first processor was turned off. Which one? There can be many variants, some of them: either these are the legs responsible for starting the second processor, or the memory block does not start from the second processor. It depends on which leg is bent.
The problem often occurs when upgrading. In this case, we are talking entirely about the human factor. However, this error is extremely common due to the fact that each model has its own installation specification. First, depending on the model, almost every server has a different number of ports. Accordingly, a certain amount of RAM needs to be added to its ports. Sometimes the task becomes complicated many times: for example, the DELL R610 model has three-channel RAM, so you need to add three strips to one channel. You can easily make a mistake by choosing from 12 slots and 4 channels.
Memory type DDR3 interface consists of three channels to two modules Rdimm or UDIMMs per channel
for single or double rank and up to two Rdimm modules per channel for rank quadra. The interface uses 2 GB, 4 GB, 8 GB, or 16 GB of Rdimm modules. 1 GB or 2 GB UDIMMs are also supported.
Therefore, one of the slats is often obtained when the memory is incorrectly calculated, or several are installed in the wrong slots, or it is not enough for an optimal configuration.
The second point, which often leads to a mismatch in RAM, is that the memory for servers is divided into power for 1.35 V and 1.5 V. The RAM for 1.35 volts starts only from the 56th version of the processor. In earlier versions of servers, you can only use strips designed for 1.5 volts. If there are different bars in the server, it will not start and will generate an appropriate error.
It looks the same way as in the case of bent sockets and the launch error of the second processor. If no error appears on the display, the server writes “Server is booting ...” and nothing appears on the screen, try checking the memory configuration. Replace the memory according to the specs and install in the correct slots to get the optimal configuration.
In fairness, I would like to finish the article by talking about the advantages of Dell servers. And they, in our opinion, completely level the problems that we described above. After all, first of all, these problems are encountered when reconfiguring servers.
We hope that our article will help you with the upgrade, diagnosis and acquisition of a used / server. Any questions? Write in the comments.
In this post we collected feedback from our engineers who have extensive experience in diagnosing, repairing and maintaining servers - each described error was solved by their sweat and blood. This information is not useful if you are serviced under warranty, but may nevertheless be useful.
Software Update
The problem when updating software is relevant for equipment DELL older than 3-5 years. Recall that all manipulations with the Dell server are made using Lifecycle, which is built into the server software, as well as the iDRAC.
')
Dell Integrated Remote Access Controller 7 (iDRAC7) is a Dell server software that connects remotely to a server, even if the server does not have power, but an internet connection and a power supply are connected. When connected to the server, the software displays information about the state of the system, and also allows you to test and update all the components: from the BIOS and ending with the drivers on the motherboard, power supplies, etc., as well as the firmware.
Here, in the standard scenario, there are two options. First, the update is in manual mode, from a flash card - in case there is no Internet connected to the server. Then go to the Lifecycle and update from a flash drive. If something does not work, we connect the server to the Internet and update, or we go through the iDRAC and restore the initial settings, install the necessary updates.
However, if the server is old enough, the problems are waiting almost immediately: when trying to install software over the Internet from the Lifecycle update and diagnostics system. The reason for this problem is that the manufacturer initially, at the dawn of this system, did not use the security protocol on the FTP server to update the firmware of the components of its servers. Now, when such certificates of protection are introduced, the old equipment of three to five years ago is sometimes difficult to update via the Internet.
The problem is solved quite simply and many of you have probably already guessed how this can be done: update the software, namely Lifecycle, but first update the IDRAC system. Updates are available for free on the official Dell technical support site. ( Check warranty and download updates to your Dell server ).
After this manipulation, you can conveniently download all the drivers and other updates via the Internet, regardless of the year of release of your server - if, of course, there is a set of updates for it on the manufacturer's website.
IDRAC update error
Recall that the iDRAC is presented in the options "standard", "express" and "enterprise". The standard comes with the purchase of the server. If we order a complete set of "express", then on the motherboard of the server we are put the appropriate licensed chip. Thus, we get a new port that allows us to receive more informative data via the Web.
With a standard license, when we connect to the iDRAC system on a work computer via a browser, we can see an overview of the system information on a standard tab, server service tag, model, and nothing more.
Via express, i.e. if we have version 7 of iDRAC, we can already update the BIOS, Lifecycle, etc. directly through the iDRAC remote access system. In earlier versions of the system, unfortunately, this feature is not provided: it can only be updated via a flash drive.
Enterprise is the most complete set of server management functions available through the iDRAC port, which allows you to fully monitor updates, a full list of components, and the current state of the server. This is a convenient and functional kit, but we managed to find in it one small but significant drawback.
After updating the version of iDRAC Enterprise 1.30.30 or 1.45.45, the system stops initializing and does not start due to incorrect settings. As a result, if we enter the iDRAC in the settings via the BIOS, we can neither turn it off, nor do anything to restore it or reboot it.
For most users, the only way out in this situation is to contact technical support. If your server is under warranty, then specialists will come and replace the motherboard. We, in turn, faced this problem and, oddly enough, a completely standard solution, described on the official Dell forum, helped: the server had to be de-energized for 5 minutes, pull out the power supplies and disconnect all the wires. True, if this does not help, then you still have to replace the motherboard.
Error E1410
Error E1410 is issued after the upgrade: you have added another processor, but for some reason the server does not start with it. Rather, it turns on, the indicators light up alternately, it makes a loud noise, but there is nothing on the screen and it writes an error on the display of the diagnostic panel.
The problem in this case is that on the socket CPU 1, the socket legs are bent, which are adjacent to the processor - either one or several at once. In the image you can see the working legs of the socket (on the left) and the place where they are bent (on the right). The legs can be manually straightened, the problem will be settled.
As you have already guessed, with this error, the second processor does not start because some of the functionality of the first processor was turned off. Which one? There can be many variants, some of them: either these are the legs responsible for starting the second processor, or the memory block does not start from the second processor. It depends on which leg is bent.
Incorrect installation of RAM
The problem often occurs when upgrading. In this case, we are talking entirely about the human factor. However, this error is extremely common due to the fact that each model has its own installation specification. First, depending on the model, almost every server has a different number of ports. Accordingly, a certain amount of RAM needs to be added to its ports. Sometimes the task becomes complicated many times: for example, the DELL R610 model has three-channel RAM, so you need to add three strips to one channel. You can easily make a mistake by choosing from 12 slots and 4 channels.
Memory type DDR3 interface consists of three channels to two modules Rdimm or UDIMMs per channel
for single or double rank and up to two Rdimm modules per channel for rank quadra. The interface uses 2 GB, 4 GB, 8 GB, or 16 GB of Rdimm modules. 1 GB or 2 GB UDIMMs are also supported.
Therefore, one of the slats is often obtained when the memory is incorrectly calculated, or several are installed in the wrong slots, or it is not enough for an optimal configuration.
The second point, which often leads to a mismatch in RAM, is that the memory for servers is divided into power for 1.35 V and 1.5 V. The RAM for 1.35 volts starts only from the 56th version of the processor. In earlier versions of servers, you can only use strips designed for 1.5 volts. If there are different bars in the server, it will not start and will generate an appropriate error.
It looks the same way as in the case of bent sockets and the launch error of the second processor. If no error appears on the display, the server writes “Server is booting ...” and nothing appears on the screen, try checking the memory configuration. Replace the memory according to the specs and install in the correct slots to get the optimal configuration.
Benefits of Dell Servers
In fairness, I would like to finish the article by talking about the advantages of Dell servers. And they, in our opinion, completely level the problems that we described above. After all, first of all, these problems are encountered when reconfiguring servers.
- Dell's "as-it-is" server is perhaps the most hassle-free and convenient for all actions with servers. This is the first thing we can say based on our experience.
- Dell also has the support of the most powerful processors at the moment: as the E3 generation and E5, up to 12-core processors (R720, R720XD).
- Dell's cold start is much faster than other big three vendors.
- Dell servers have a large selection of baskets for connecting HDDs - 2.5 inches and 3.5 inches, in which you can connect up to 12 HDDs (in the case of the R720XD). With adapters, you can connect SSD drives. At 3.5 inches, there are baskets for 4, 6, 8, 12 discs, and for 2.5 inches - up to 8.
- Almost all Dell models have the ability to connect redundant power supplies.
- Technical support for applications responds very quickly.
- Therefore, despite the fact that Dell servers are a bit more expensive than analogues from other suppliers, the investment in these servers is worth it.
We hope that our article will help you with the upgrade, diagnosis and acquisition of a used / server. Any questions? Write in the comments.
Source: https://habr.com/ru/post/270091/
All Articles