Monday, October 31, 2011
Recently, I've been working on a project which involves remote monitoring and configuration of server and network devices. Obviously, my research has led me to further discover the intricacies of the SNMP protocol and agents which can often provide extensive monitoring and configuration capabilities.
I've obviously been looking at native and third-party, proprietary and open source, SNMP agents for Windows and Linux. No doubt, configuration parameters can vary drastically depending on the information and level of configurability that you desire but they all rely on similar concepts and use similar vocabulary. These concepts are best outlined in the following locations.
The best example of its structure is probably represented by the following picture.
It will most likely remind you of postal addressing systems, directory service hierarchies and other hierarchal structures such as those used by Domain Service System (DNS) with each level being represented by numbers and/or shorthand strings. At the end of the hierarchy is a string/counter/number which represents the value of the concerned attribute of the piece of hardware and/or software in question at any point in time. These attributes can range from the name of the installed Operating System, to the total number of bytes sent on a particular Network Interface Card (NIC).
Based on the intent on my project it seems clearer that using SNMP as a basis for monitoring may be overkill (though I'll add some SNMP functionality). While it allows for a more finer grained image of the situation, it is also clear that much of this information is redundant especially when you are using 'generic hardware' (whether it is server and or desktop class). As such, it becomes more obvious the reasoning behind extensive templates that often accompany more mature monitoring systems such as ZenOSS, Zabbix, and Dell OpenManage.
While it is clear that SNMP is a useful and mature protocol there have clearly been moves to modernise it (for instance, through enhanced security, accessibilty, and configurability) and there are alternative technologies such as Link Layer Discovery Protocol (LLDP) and Cisco Discovery Protocol (CDP).
Friday, October 21, 2011
Sometimes you'll encounter a continuing problem over and over again. One of the ones that has come to light for me has been related to discrete NVIDIA graphics chip on laptops (across multiple brands believe it or not including HP, Compaq, Dell, and Apple) which have revolved around excess heat not being dispersed adequately which means that over time the graphics chip can come away from the logic board. A lot of the time this results in an unbootable laptop with the following symptons, power and LED indicator lights turn on accompanied by a black screen. Even though there has been a partial recall of affected laptops there are many 'out in the wild' which still have the same problem.
While conventional solder and mechanical pressure can be used as a means of providing a medium to long term fix on traditionally packaged and attached chips this is not the case with Ball Grid Array (BGA) graphics chips which seems to be a more popular methodology moving forward.
However, the probem is when you have mounting issues they aren't so easy to fix if you don't have the required, proper equipment required in order to complete the job. Moreover, often taking it to a 'professional' may cost the same amount as a new laptop/logic board. Numerous accounts online of this particular problem and some creative and unusual solutions have been proposed that are more 'cost effective' (but are also shorter term fixes). Some of these include:
- stripping down the machine down to the logic board and then 're-flowing' it by putting it in a convection oven or even using direct heat such as using a butane torch
- wrapping it in insulating material, turning it on and using the laptop's own heat in order to 're-flow' the connections
- using direct heat and a BGA kit in order to remove the damaged chip from the logic board, the 'balls', and then using the included solder balls (often of higher quality than the original) to fix problem areas
However, one thing I have been considering is whether or not it is feasible to attempt another solution. Namely, creating the solder balls themselves (too difficult on a large scale), or else using solder wick in order to clean up damaged BGA balls and then using flux gel in order to crudely create 'balls' between the graphics chip and the logic board. While I have been able to partially resurrect one board with this method I'll need more time to determine whether or not whether it is a complete long term fix...
Thursday, October 20, 2011
The following represents some notes that I compiled while researching VERITAS Backup Exec (a commercial backup software solution that has since been purchased by Symantec). Obviously, these notes are based on an older version of Backup Exec so use your discretion.
Intro to Backup
SLA - Service Level Agreement
Full - complete copy of all files
Incremental - only files changed since last backup
Differential - all files changed since last full backup
Online Backup - everything accessible all the time
Nearline Backup - data from backup periodically archived to secondary backup server. Requires minimal effort to restore backup to main
Offline Backup - files unavailable during backup
Disaster Recovery Technologies : Hardware
Tape - linear, cheap, high storage capacity density
Hard Drives - less mobile, medium price per unit storage capacity,
Removable Disks - portable
Optical Media - cheap, portable,
NAS - relatively inexpensive now, RAID
SAN - can be expensive
Disaster Recovery Technologies : Software
Server-based Backup - server/administrator determines backup configuration
Client-based Backup - client/users determines backup configuration
Frozen Image Backup - live snapshot of files possible
SAN-based Backup - backup to SAN
LAN-free Backup - typically uses a SAN to backup using a dedicated Fibre connection
Server-free Backup - client system/software initiates the backup. Stored to tape/SAN
Applying Technologies : Client-based system
Backup Program on Each Machine
Backup client on each machine with a single, central backup server
Backup Program on Each Machine - often OS includes basic backup capabilities. Free options available but they often from lack of support, flexibility, and robustness
Client Controlled on a Network - user configuration, schedule, backup server stores on tape. User configuration uncommon because users want to backup everything, most users aren't technially aware eough, and most administrators want a consistent environment
Backup Jobs - sometimes also called policies
Generic Troubleshooting - hardware and/or software issues, test partial/complete backup and restores
Applying Technologies : Network-based System
WAN/LAN - primary difference is physical distance between systems
Server-based Solutions - most are server based with a administration utility
Heterogenous Environments - most commercial solutions now have the ability to operate across multiple platforms by using an 'agent'
Replication - latest copy and transparency during failure but server/bandwith requirements are greater along with extra resources for extra hardware/server. Run-Time Vs Traditional Vs Frozen Image backup
Scheduling Issues - timezones, resource consumption, mutual exclusion
Networking Utilization Issues - schedule during off hours, hard to schedule in a genuine 24/7 environment though
SAN Solutions - cost, compatiblity, configuration
Tape Sharing - shared tape drives, tape libraries, robotic libraries
Tape Rotation - saves from wear, space by not having irrelevant backups; schemes include daily, GFS (Grandfather-Father-Son), Tower of Hanoi
Looking Ahead in Backup Technology - increased sophistication/capacity of tape solutions, maturation of disk solutions, NAS/SAN technologies continue to evolve, new technologies will come into existance
Introduction to Backup Exec
Single and Multi-Server Editions
Remote Agents - best to check support for your OS prior to evaluating/purchasing
Autoloader/Robotic Library Support
Library Expansion Option (LEO)
Advanced Open File Option (AOFO)
Backup Options for Various Databases - MS Exchange/SQL/Sharepoint, Oracle, Lotus Domino, SAP
Intelligent Disaster Recovery (IDR) - known as bare metal restore now, involves a scripted installation of a minimal OS and then restoring to some initial configuration/restoration point from there
Shared Storage Option (SSO) - used by various servers/systems to access the same medium
Backup Exec Installation : Suggestion and Pitfalls
Install latest Service Pack/Patches
Update Microsoft Data Access Components (MDAC) - OLEDB32.dll
Update Microsoft Installer (MSI) - MSI.DLL
Check HCL/System Requirements
Try to have dedicated chain/channel for the storage/backup medium
Security Considerations - use the account that was automatically created, don't try to customise it as it can result in strange permission errors
Getting Your Hardware to Get Along with Backup Exec
Drive Pools - drives from the same server
Cascaded Drive Pools - used to deal with when normal drives can't deal with size of backup. Responsibilty of backups fall to these particular drives if the normal drives pools are unable to
Hardware Configuration Problems - determine whether it is a program, OS, BIOS issue. Use the built-in Wizard to try solve other issues
ADAMM (Advanced Device and Media Management) Log
Install drivers (all drivers in one package) from website if necessary, http://support.veritas.com
BEUTILITY.EXE - to reset hardware tracking databases, BEDB.MDF/BE_DLO.MDF
Database Catalogue - backup jobs, media set information, hardware statistics, automatically backed up after 24 hours of programs/services running
Physical Issues - BIOS, Power Cycle (server and media drives), physical connections
Overwrite Protection - to stop media from being overwritten prematurely
Append Period - stops any data from being appended to media at all as opposed to overwritten
Media Sets - set of media based on overwrite/append settings
Time-to-Recovery - balance between time/convenience/cost
Granularity - frequency of backups
Backup and Restore
Shadow Copy Components - System State/User Data/Service State
Remote Selections - will remove redundant data from backup selection by default
The Remote Agent for Windows Servers (RAWS) - must be installed on remote server to complete remote backup, can be installed by going to Tools > Serial Numbers and Installation or running setupaa.cmd
The Advanced Open File Option (AOFO) - must remove Remote Agent if already installed otherwise just install AOFO and it will install RAWS as well
Backup Performance - Avoiding Bottlenecks
Advanced Backup Options
Pre- and Post-Job Commands - run custom scripts
IP-only backup - as opposed to NETBIOS addressing/resolution
Backup Folders - backup to an arbitrary folder (NAS/SAN) rather than media
Hierarchical Storage Management (HSM)
An Introduction to Using Backup Exec with Other Platforms
Microsoft Exchange - allows for "brick level" backup of mailboxes
Microsoft Outlook - need to install MAPI compatible email client on media server (normally Outlook). Backup Exec logs in each user individually and downloads all emails to allow for "brick level" backup
Information Store - Backup Exec uses ESEBCLCLI2.DLL or EDBBCLI.DLL (Exchange 5.5) to communicate with Exchange
Types of Exchange Database Backups - Full/Database&Logs (flush committed logs), Incremental/Logs (flush committed logs), Differential/only transaction logs, Copy/same as Full (none of the transaction logs are deleted though)
Microsoft SQL Server - SQL Enterprise Manager must be installed on media server.
Types of SQL Database Backups - Database (backup everything along with transaction logs), Log (backup only transaction logs and delete old transaction logs from database), Log No Truncate (same as Log but does not delete old transaction logs from database), Consistency Check Before Backup (check database consistency before backup. This may cause a performance hit on a high transaction/traffic server though)
UNIX and Linux - be_agnt.tar
This is a continuation of my other post: http://dtbnguyen.blogspot.com/2017/04/life-in-vietnam-prophetspre-cogs-12-and.html - more Yo...
- many people are often critical of choices that are often made in other countries. If you think carefully about it, massive loops have bee...
If you're like many others on the planet you probably can't be bothered purchasing a new piece of software in order to test an appl...