Search Results

Search found 14111 results on 565 pages for 'virtual machines'.

Page 434/565 | < Previous Page | 430 431 432 433 434 435 436 437 438 439 440 441  | Next Page >

  • Storing Windows Updates for reuse

    - by Saiyine
    At work we update lots of computers using Windows Update. Windows XP and 7 all day long, rarely some Vista. We do it through a corporate proxy, as connecting them to a domain to build a Wsus server is out of the question, so we download about two gigabytes a day of the very same updates everyday. I've tried WSUS Offline. It's pretty complete but when it finishes it's common to be still missing hundreds of megabytes of updates, because its intention is not to fully update a system but to install the critical updates, as the developers explain in the forums. Now I'm trying with Autoupdater. It's far worse, with poor capabilities for non-English Windows XP, but at least it gives the option to install non critical updates in Windows 7. It still misses hundres of megabytes of updates after fully updating the system. And finally, both doesn't install the driver related updates of Windows 7, so they at most save us a couple of hundreds of megabytes and a reboot (with the associated login to the computer and to the proxy) out of three or four. So, is it possible to somewhat extract the installed updates in a Windows 7 system and not having to download the same updates again and again at least with machines with the same hardware? Or even better, a generic package with all the updates?

    Read the article

  • Datacentre Rack naming convention with flexibility for reassignment of server roles

    - by g18c
    We are just shifting across to a new rack and until now have used names of cartoon characters. This is not going to work anymore, and need a better naming convention. Physically i would like to name the servers by location, and then have an alias as to its actual function/customer, i.e. Physical name LONS1R1SVR1 meaning London, suite 1, rack 1, server 1 Customer Alias Since the servers can be reassigned from time to time, for the above physical server name, i would have an alias as a column in a spreadsheet, that would be set to the customers host-name, i.e. wwww.customerserver1.com Patching For patching, I am looking at labeling up the physically connections, i.e. LON1S1R1SVR1-PWR1 LON1S1R1SVR1-PWR2 LON1S1R1SVR1-ETH0 LON1S1R1SVR1-KVM Ultimately if i am labeling cables, I really want to avoid putting LON1S1R1SQLSVR on any patch cord in case the server gets formatted and changed from a SQL server to a WWW server which would need to relabel all the patch cords also. In addition, throwing in virtual machines, i have got confused very quickly. I appreciated that it may be confusing having a physical host-name and customer alias. Please let me know what you run with and any other standards or best practices that i can follow?

    Read the article

  • How to turn off Windows Azure's "This copy of Windows is not genuine" message?

    - by Sid
    Is there any setting/configuration item to avoid Windows Azure from printing that error on the screen or detecting it? I've put a screenshot below that shows the message when you RDP into the web role. My web role runs on Windows Azure Guest OS 1.17 (variant of Windows Server 2008 SP2) Background: I was explaining our architecture to some outside engineers (NDA'd and all) and had to demystify the webrole as they were unfamiliar with Azure. I RDP'd into the VMs running the Web Role when one of their engineers gasped "are you guys running pirated copies of Windows in the cloud?" I also noticed that within the RDP screen, the Azure machines had "This copy of Windows is not genuine" on the bottom left corner. Now obviously, Microsoft is running their own OS in their own datacenter with no influence from me. So no 'piracy' here, despite that obvious warning. However, they seemed so distracted by this ("how can it be? really? hmmm?") that we wasted more time talking about it than the actual matter at work. Like I said, they have little exposure to Azure but have value add elsewhere. I want to get rid of this so I don't have to explain this in the future. PS Microsoft: If you're going to modify Windows Server <XYZ> into Windows Azure <A.B> , you should also modify the code that verifies product integrity.

    Read the article

  • download and process a file by ftp at set intervals, with error handling, rescheduling and status messages

    - by compound eye
    I want to download a data file from a remote ftp server to my machine at regular intervals. Once the file is downloaded I want to call another script which will process the file. My development machine is mac os x, the eventual deployment environment is linux. What's would be the stock standard way to automate this? I know I can use cron to schedule curl to download and to run a script that will process the downloaded file at regular intervals, and I know could write a slightly more complex script or an application that would do this and add error handling, rescheduling and sending status emails. But one of my requirements for this project is to write as little custom code as possible, instead I should try to use standard, tried and true existing tools, and if I do have to write code, to try and write the most straightforward code possible. The reason for this is the code will potentially be installed on a large number of machines, all of which will need to be tweaked, customised and maintained by different people, long after I am gone from the project, so the intention is to use well documented, well supported tools as much as possible. This seems such a common task, there must be tools and scripts all over the internet, written by people who have carefully considered everything that could possibly go wrong when you need to download and process a file from a remote server at regular intervals, with error handling, rescheduling and sending status messages. Is that what Expect is for? What would you recommend? (the system will be downloading weather prediction data every six hours, so that the system can prepare in the event of bad weather warnings)

    Read the article

  • Problems with USB-Devices using VDR

    - by emmsinator
    Hey Guys, I'm using VDR on vSphere4. It works sucessfully. I've already backuped several VMs with VDR and I like it very much. But now we got a problem. We have 2 VMs, using an USB-Device Server with a stick plugged in, which is definetely need by these 2 VMs for Licensing and so. Every time, I start the Backup process, the VMs lost the communication to the USB-Server and its stick after building the snapshot and while online. Because of that, the software on these VMs can't work correctly. I have to restart both Machines to solve this problem. These fact is bad for an automatic backup. Does VDR have a special function for those cases or is something like this already known? It would be no problem, to shutdown the servers for building snapshots on Saturday or Sunday. Can VDR initiate a shutdown before starting the backup process? Otherwise I must try to use scripts, but that wouldn't be so nice. Thanks a lot for your help.

    Read the article

  • IDE compatability with SATA image

    - by Ormis
    We had an old CNC machine's hard-drive fail recently. The hard-drive is an old 1275MB IDE (Seagate) and there were defiantly bad sectors on it. I was able to image the contents of the drive onto a drive in my computer before it became completely unusable (I used DD, replacing all bad sectors w/ 0s). After running a couple chdsks, the SATA drive will boot off of the image. This is great, but there's one problem. The CNC machine old and requires IDE, I've attempted to copy the currently booting image off of the SATA drive and onto IDE drives numerous times in numerous ways and every time I do so the machines return that a boot device cannot be found. Some other information: The file system is fat32, running windows 98 The SATA drive is an 80gb drive I have tried copying the image to three 20gb and two 80gb IDE drives I have checked the jumper on the back of the IDE drives when using them If anyone has any ideas, questions, suggestions, etc. please let me know. P.S. I would just put a fresh install of win98 on the machine if i had the installation media (so that's out of the question). And if it comes to it, this is my last week working here, so I'll leave that to my co-worker. EDIT: Also, I have tried using Clonezilla as well as straight up DD to copy the image to the IDE drives.

    Read the article

  • Windows Server 2012 licensing issue preventing RDP connections?

    - by QF_Developer
    I am witnessing an unusual behaviour on 1 of 5 Windows Server 2012 R2 machines (clean install) that is preventing any remote connections from being established via RDP. I have run through the prerequisites for RDP here but I am finding that any remote connection attempt instantly stops the "Windows Protection Service". When I check the event logs I see the following entry. The Software Protection Service has stopped Event ID: 903 Source: Security-SPP From what I have read Security-SPP is tasked with enforcing activation and licensing, it appears that RDP requires this service to be in the running state. Is it possible that I have inadvertently activated this instance of Windows with a key that has already been associated to another instance (We have 5 keys as part of an MSDN subscription)? Would this be sufficient to block RDP access? When I look under System Properties (Windows Activation) it states that Windows is activated and there are no other obvious indicators that there's a licensing issue. EDIT 1: I ran a Powershell script to display the product keys for all servers in order to check for any duplication. For the problematic server I am getting the message The RPC server is unavailable.

    Read the article

  • 100% uptime for a web application

    - by Chris Lively
    We received an interesting "requirement" from a client today. They want 100% uptime with off-site failover on a web application. From our web application's viewpoint, this isn't an issue. It was designed to be able to scale out across multiple database servers, etc. However, from a networking issue I just can't seem to figure out how to make it work. In a nutshell, the application will live on servers within the client's network. It is accessed by both internal and external people. They want us to maintain an off-site copy of the system that in the event of a serious failure at their premises would immediately pick up and take over. Now we know there is absolutely no way to resolve it for internal people (carrier pigeon?), but they want the external users to not even notice. Quite frankly, I haven't the foggiest idea of how this might be possible. It seems that if they lose Internet connectivity then we would have to do a DNS change to forward traffic to the external machines... Which, of course, takes time. Ideas? UPDATE I had a discussion with the client today and they clarified on the issue. They stuck by the 100% number, saying the application should stay active even in the event of a flood. However, that requirement only kicks in if we host it for them. They said they would handle the uptime requirement if the application lives entirely on their servers. You can guess my response.

    Read the article

  • Windows disk change monitoring for malware analysis

    - by SuperDuck
    Not sure if this question belongs to here, because it has some relations with 'serverfault' (system backups) and 'stackoverflow' (software analysis). I'm looking for a solution to monitor disk changes on a Windows system and selectively revert them. It should be able to handle live files like registry parts, so may need to be an offline backup software. It shouldn't silently pass over files which the current admin user doesn't have permissions on (files with no permission entries or owned by the 'system' user) Registry change tracking would be a bonus but is not a requirement I use virtual machines for malware analysis, there is even no solution to list file changes in disk snapshot files (delta VMDK). I currently use Ashampoo for monitoring changes. Though it's the best one between similars, it's not a good software and hasn't really evolved in many 'platinum', 'deluxe' versions released in the last 10 years (it even used non-resizable windows until the latest version). The real problem is it misses some disk / registry changes. Perhaps it only compares modification dates and doesn't catch a change if the dates are preserved. So, I think the solution should compare files using hashes, or file sizes at least. There are numerous backup software out there and I'm sure one can handle this, offline or online.

    Read the article

  • (Win7) Gets stuck with ~1% CPU. Especially with multithreading

    - by meow
    Windows 7 32 bit, up to date, Intel i7 860. (For some reason the company runs 32bit Windows everywhere.) I tried to update all motherboard drivers etc. as far as possible. I have a performance issue with a machine which appears in connection with multithreading (or so I think). As an example (and where I most often see it, but it appears on other programs as well): ProteoWizard is a file conversion tool for mass spectrometry files. I can add a list of files and it will attempt to process up to 8 files in parallel (quadcore x 2 threads/core). If I choose 1 to 6 files, I start the process and it goes straight through. If I have =7 files in the queue, conversion goes to ~20%, then gets stuck for 15 seconds, then continues again, always in "chunks" of a few % before getting stuck again. During the time the process is stuck, CPU is at 1%. RAM is not limiting, it is maybe at 70% or so and not going up. I don't get the same problem on other, even slower machines. The computer gets also stuck at 1% CPU doing nothing on other occasions, but for multithreading it is most frequent. Where should I look for the problem?

    Read the article

  • Apache Server with memcache, varnish and php slow request times

    - by coolestdude1
    My issue is that these servers are taking rather long for request about 2 seconds on average just to serve files. When we had just one server doing everything it was noticeably faster even with the same web app (Drupal 6 and Drupal 7). I want to get this number down to a reasonable level and so I need some help getting to the bottom of why the request times are so slow. This can cause the webapp to hang on post or put and generally leads to a bad user experience on my sites. PS: I am more of a server newbie so this has confounded me for quite some time. The domains: collabornation.net nptrainingworks.com (they run off the same two webservers using vhost configs) The Gear: Two Rackspace 4 Gig servers running CentOS 6.2 Final They have a mounted file system (gluster) that is used to keep files the same on both machines. They are behind a rackspace load balancer running round robin. Mysql is run using php-pdo and php-mysql as such mysql is run on another instance running memcache on that machine with phpMyAdmin located there as well. Apache version number 2.2.15-15.el6.centos.1 (httpd.x86_64) Varnish version number 3.0.2-1.el5 (varnish.x86_64) PHP version number 5.3.14-1.el6.remi (php.x86_64) Configs Linked Below Apache Conf Vhost Conf Varnish Backends Varnish Defaults Varnish Acl PHP INI Again need some help, much appreciated!

    Read the article

  • computer fails to boot during/after POST for five or six boots, then works

    - by N13
    For the last few days, my computer has had issues booting. I've seen two different behaviors: The screen displays the graphics card information, then begins to list the RAM, hard drives, etc. At different points in this process (after the graphics info), the computer shuts off. After five or six attempts, it then boots normally. In roughly the same time frame, the computer freezes, and fails to boot. I think it boots successfully on the next attempt. I've also noticed that in some instances, the computer freezes on shutdown. It gets right to the point where it should shut off, but doesn't. I recently combined the best parts of two different machines into this one. I'm booting to GRUB, with Ubuntu 12.04, Linux Mint 11 and Windows Vista (unfortunately) as my OS options. It has an Enermax Modu82+ 525W power supply, and I've used an online calculator to determine that my load shouldn't exceed 400W. I even unplugged a hard drive, but that didn't help. I found the latest BIOS, patched it and checked the settings, but that didn't fix it. I'm fairly certain this issue didn't exist at first, but might have started when the power at my new apartment dropped for a second. The machine is plugged into a surge protector strip, but it's old and I've heard they lose effectiveness with age. Is a power dip as damaging as a spike? If something were damaged, why would it boot successfully after five or six attempts? It's almost like the BIOS or PSU need to be primed. The trouble with debugging is that there seems to be a "grace period" after shutdown where the issue doesn't present itself again. What should I try next?

    Read the article

  • Assigning cores to VM in vSphere

    - by user114933
    Complete vSphere newbie here... Background: So, I have a 12 core machine with 24 VMs on it. Currently, all the processing power is shared between these VMs equally. The question: Can I configure one VM to be given two CPU's worth processing no matter what's happening on the other machines? My Research: I tried two things in vSphere... I set the reservation and limit on one VM to equal the same as two cores. To test if my objective was being reached, I measured the time it would take to gzip a file when other VMs were running nothing and when other VMs were running CPU intensive operations. I expected the time to gzip the file would be the same because this VM gets priority for some processing. Unfortunately, the time taken to gzip the file when other VMs were running something was significantly more than when other VMs were not running anything. I tried setting the Hyperthreaded Core Sharing mode to Internal hoping that this would mean that my VM would get at least an entire core to itself. This did not work either. Thanks in advance!

    Read the article

  • TortoiseSVN client slows Explorer to a crawl in Windows XP running in Parallels

    - by Cory Larson
    I thought I'd make my first SuperUser question relatively simple, though it's the kind of question that may not get many responses as I'm not directly involved with the issue. A colleague does his development in Windows XP running in Parallels on his Mac. We've just migrated our VSS repository to SVN, and we've gone with TortoiseSVN as our client of choice with the Ankhsvn plugin for Visual Studio. On his XP instance, after installing TortoiseSVN, browsing through folders using Explorer is extremely slow; about 15 - 30 seconds before the contents of the next folder displays. It's the slowest when opening My Computer. Once he reaches a folder that contains the working content of an SVN project, Explorer behaves quickly again as expected. It seems that TortoiseSVN may be spending a bunch of time searching subfolders for stuff so it can do its icon-overlay thing, but that's just a guess. I've used TortoiseSVN for years on both XP and Vista on far less powerful machines without any issues with Explorer, so I'm attributing the slowness to it being run in a VM, though that may not be the actual issue. So has anyone encountered similar performance issues, and/or know of a fix? Keep in mind that any requests to make changes to his configuration will need to be communicated and thus my response time might be slow. Thanks everyone!

    Read the article

  • Simple Distributed Disconnected way to sync a directory

    - by Rory
    I want to start regularly backup my home directory on my ubuntu laptop, machine X. Suppose I have access to 2 different remote (linux) servers that I can backup to, machines A & B. Machine X will be the master, and should be synced to A and B. I could just regularly run rsync from X to A and then from X to B. That's all I need. However I'm curious if there's a more bandwidth effecient, and hence faster way to do it. Assuming X is going to be on residential style broadband lines, and since I don't want to soak up the bandwidth, I would limit the transfer from X. A and B will be on all the time, however X, will not be, so I'd also like to reduce the amount of time that X is transfering, potentially allowing A and B to spend more time transfering. Also, X won't be connected all the time. What's the best way to do this? rsync from X to A, then from A to B? Timing that right could be troublesome. I don't want to keep old files around, so if I was to rsync, then the --del option would be used. Could that mean something might get tranfered from A to B, then deleted from B, then transfered from A to B again? That's suboptimal. I know there are fancy distributed filesystems like gluster, but I think that's overkill in this case, and might not fit with the disconnected nature.

    Read the article

  • NVidia TwinView - slow rendering on dual desktop [closed]

    - by lisak
    Hey, does anybody have experience with it ? I've set it up 4 times on 4 different machines. And there was always problems with slow rendering ( for instance : scrolling pages in browser is not fluent). But there always was something that finally made it work perfectly... I remember that one time this option helped, but not now Option "RenderAccel" "1" Nvidia geforce 8400GS or Zotac geforce 9500GT Monitors connected via dvi and hdmi connectors proper nvidia driver installed Section "ServerLayout" Identifier "X.org Configured" Screen 0 "Screen0" 0 0 InputDevice "Mouse0" "CorePointer" InputDevice "Keyboard0" "CoreKeyboard" Option "Xinerama" "0" EndSection Section "Files" ModulePath "/usr/lib64/xorg/modules" FontPath "/usr/share/fonts/local" FontPath "/usr/share/fonts/TTF" FontPath "/usr/share/fonts/OTF" FontPath "/usr/share/fonts/Type1" FontPath "/usr/share/fonts/misc" FontPath "/usr/share/fonts/CID" FontPath "/usr/share/fonts/75dpi/:unscaled" FontPath "/usr/share/fonts/100dpi/:unscaled" FontPath "/usr/share/fonts/75dpi" FontPath "/usr/share/fonts/100dpi" FontPath "/usr/share/fonts/cyrillic" EndSection Section "Module" Load "dri2" Load "glx" Load "extmod" Load "record" Load "dbe" EndSection Section "InputDevice" Identifier "Keyboard0" Driver "kbd" EndSection Section "InputDevice" Identifier "Mouse0" Driver "mouse" Option "Protocol" "auto" Option "Device" "/dev/input/mice" Option "ZAxisMapping" "4 5 6 7" EndSection Section "Monitor" Identifier "Monitor0" VendorName "Unknown" ModelName "Acer AL1715" HorizSync 30.0 - 83.0 VertRefresh 50.0 - 75.0 EndSection Section "Device" Identifier "Nvidia" Driver "nvidia" VendorName "NVIDIA Corporation" BoardName "MSI big bang-fuzion" EndSection Section "Device" Identifier "Device0" Driver "nvidia" VendorName "NVIDIA Corporation" BoardName "GeForce 8400 GS" EndSection Section "Screen" Identifier "Screen0" Device "Device0" Monitor "Monitor0" DefaultDepth 24 Option "RenderAccel" "1" Option "AllowGLXWithComposite" "1" Option "TwinView" "1" Option "TwinViewXineramaInfoOrder" "DFP-1" Option "metamodes" "CRT: 1280x1024 +1920+0, DFP: 1920x1080 +0+0" SubSection "Display" Depth 24 EndSubSection EndSection

    Read the article

  • VMWare Server modifying files related to paused VMs, is this expected?

    - by David Spillett
    While refreshing the backup of a VM used for testing, I experienced the following warning from tar: tar: /VMsR0/cli_noddyco_test/VM2K8_32_web.vmem: file changed as we read it The VMs in question were paused at the time. My first though was that I'd mixed up the machines and was trying to backup something that was still actively running. To be sure I unpaused and properly shut down the VM, and the vmem files that tar reported changing vanished as I would expect. Is it normal for VMWare Server to touch or alter files for paused VMs like this, or is there likely something amiss with our setup? If this is expected behaviour, is just touching the vmem file (and so altering the last modification date without actually changing content)? If it is normal for files relating to paused VMs to be updated I shall have to revise our backup procedures to make sure the VMs are fully shut down fully rather than just pausing them (this isn't a problem, but it seems strange and I'd prefer to understand what VMWare is doing and why instead of just dismissing it as "one of those things" and working around it). For further detail: the host in question is VMWare Server version 2.0.2 running on 64-bit Debian/Lenny, and that VM did not have a snapshots at the time. We have backed up paused VMs this way in the past with no such warnings from tar.

    Read the article

  • Advice for UPC/Surge Protector in home office

    - by Fred
    I'm just starting out as an independant developer, mostly Unix stuff with some Windows thrown in occasionally. I've been running two machines, a linux and a windows dev machine. Long story short, we had a bad storm come through last week and I unplugged one machine, forgot to unplug the other and the p/s and mobo ended up dead. Luckily I backup to an external service religiously (rsync.net for anyone interested), so there was no loss in data, but it did show me a glaring hole in my current setup, namely, lack of UPC and Surge Protection (this has honestly never been an issue before). Can anyone recommend a UPC/Surge Protector for a home office? It only needs to support a single machine (I opted to use vmware instead of rebuilding that machine), but it's a quad core Phenom 2 with a 1k watt p/s. This is outside my experience so I thought I'd get some input from others. I'm looking for something that's reasonably priced and does the job reasonably well. I don't need absolute 100% uptime, just something to protect my PC better than it is now.

    Read the article

  • Simple, centralized user management on a small LAN - NIS or LDAP?

    - by einpoklum
    I'm setting up a small LAN for my team. It will, for all intents and purposes, not be connected to any external networks. I would it to have centralized control of user accounts (at least, I think I'd like that; I'm also considering using puppet, so theoretically I could just push /etc/passwd changes, or something). The number of machines is fixed, but not very small. Mostly they're 'attached' to a single user, but sometimes people work remotely on someone else's box; and there are a couple of servers. I've read this question, but my scenario is much simpler (even simpler than in this question) and I'd like to do something (relatively) quick, with not much hassle, but not a dirty totally-insecure hack. Is NIS relevant for my scenario? If not, what's the most hassle-free way to set up LDAP (or LDAP+Kerberos) to achieve the same? Notes: I have no experience with setting up either NIS or LDAP. We use Debian-flavored Linux distributions, mainly Kubuntu 12.04 (not my choice, but that's the way it is).

    Read the article

  • Screen randomly goes blue/black/white

    - by FubsyGamer
    Problem Randomly, while using my computer, the monitor goes dark grey/almost black, or it goes white with faint grey vertical lines, or it goes blue with black vertical lines. It's as if the computer powers off. People tell me I sign out of Skype, Spotify stops playing when it happens, etc. When I look at the tower, it doesn't seem like it's off at all. Nothing changes, fans are spinning, lights are on, etc. If you were only looking at the tower, you'd never know there was a problem The only way I can get it to come back up is to push and hold the power button and turn it off, then back on This never happens while I'm playing video games. I've done 5-6 hour sessions of League of Legends, and it doesn't do anything When I'm just browsing the web, reading email, checking Reddit, etc, it happens all the time. It can happen multiple times in a session, it usually takes only about 5 minutes from the time I start browsing to when the computer crashes This started happening after I moved to a new apartment (this has to be relevant somehow, it was not happening where I lived before) There is nothing in the crash logs or event logs System Specs i5 2500k CPU AMD Radeon 6800 GPU Gigabyte z68a-d3h-b3 motherboard WD VelociRaptor 1 TB HDD Screenshots Device manager About screen Things I have tried I was getting a WMI Error in my event logs, but I fixed it using Microsoft's fix, KB 2545227 I was using Windows 8. I wiped the HDD and downgraded to Windows 7 64 bit I took out the video card and used a can of air to totally clean out the video card, all fans, and the inside of the computer in general. I made sure all of the video card pins were fine, then reconnected it I tried to update my motherboard BIOS, but anything I downloaded from Gigabyte was only for 32 bit machines, not 64. I don't even know how to tell what my motherboard BIOS is at right now I am using a power strip, and anything else connected to it works just fine If I re-seat the monitor cable while this is happening, nothing changes Please, help me. I've been battling this for several weeks now, and it's so frustrating it makes me not even want to use the computer.

    Read the article

  • SSH Keys Authentication keeps asking for password

    - by Rhyuk
    Im trying to set access from ServerA(SunOS) to ServerB(Some custom Linux with Keyboard Interactive login) with SSH Keys. As a proof of concept I was able to do it between 2 virtual machines. Now in my real life scenario it isnt working. I created the keys in ServerA, copied them to ServerB, chmod'd .ssh folders to 700 on both ServerA,B. Here is the log of what I get. debug1: SSH2_MSG_KEXINIT sent debug1: SSH2_MSG_KEXINIT received debug1: kex: server->client aes128-ctr hmac-md5 none debug1: kex: client->server aes128-ctr hmac-md5 none debug1: Peer sent proposed langtags, ctos: debug1: Peer sent proposed langtags, stoc: debug1: We proposed langtags, ctos: en-US debug1: We proposed langtags, stoc: en-US debug1: SSH2_MSG_KEX_DH_GEX_REQUEST sent debug1: expecting SSH2_MSG_KEX_DH_GEX_GROUP debug1: dh_gen_key: priv key bits set: 125/256 debug1: bits set: 1039/2048 debug1: SSH2_MSG_KEX_DH_GEX_INIT sent debug1: expecting SSH2_MSG_KEX_DH_GEX_REPLY debug1: Host 'XXX.XXX.XXX.XXX' is known and matches the RSA host key. debug1: Found key in /XXX/.ssh/known_hosts:1 debug1: bits set: 1061/2048 debug1: ssh_rsa_verify: signature correct debug1: newkeys: mode 1 debug1: set_newkeys: setting new keys for 'out' mode debug1: SSH2_MSG_NEWKEYS sent debug1: expecting SSH2_MSG_NEWKEYS debug1: newkeys: mode 0 debug1: set_newkeys: setting new keys for 'in' mode debug1: SSH2_MSG_NEWKEYS received debug1: done: ssh_kex2. debug1: send SSH2_MSG_SERVICE_REQUEST debug1: got SSH2_MSG_SERVICE_ACCEPT debug1: Authentications that can continue: publickey,keyboard-interactive debug1: Next authentication method: publickey debug1: Trying private key: /XXXX/.ssh/identity debug1: Trying public key: /xxx/.ssh/id_rsa debug1: Authentications that can continue: publickey,keyboard-interactive debug1: Trying private key: /xxx/.ssh/id_dsa debug1: Next authentication method: keyboard-interactive Password: Password: ServerB has pretty limited actions since its a custom propietary linux. What could be happening? EDIT WITH ANSWER: Problem was that I didnt have those settings enabled in the sshd_config (Refer to accepted answer) AND that while pasting the key from ServerA to ServerB it would interpret the key as 3 separate lines. What I did was, in case you cant use ssh-copy-id like I couldnt. Paste the first line of your key in your "ServerB" authorized_keys file WITHOUT the last 2 characters, then type yourself the missing characters from line 1 and the first one from line 2, this will prevent adding a "new line" between the first and second line of the key. Repeat with the 3d line.

    Read the article

  • Make a drive from one machine appear as a physical disk in another machine.

    - by Roberto Sebestyen
    I want to take a physical disk (or part of a disk) in one machine (call it machine-A) and I want to make it available in another machine (machine-B). But I don't want to map a network drive. I want it to appear in machine-B as a physical drive. Even though it is not a physical drive. The reason I want to do this is i want the ability to create shares in machine-B on that drive. Since I cannot do that on mapped drives, I need to use some utility that fools machine-B to think that it is a physical drive, and treat it as such. Both of these machines are windows server 2003. I heard about NFS, It sounds like what could be the solution to my problem. But isn't that a Linux/Unix protocol? What tools can I use to make this happen? Are there any open source solutions? I don't care what the solution is, as long as it achieves the end result, preferably open source solution though. Thanks for reading guys and gals!

    Read the article

  • /etc/hosts: What is loghost? (fresh install of Solaris 10 update 9)

    - by cjavapro
    # # Internet host table # ::1 localhost 127.0.0.1 localhost XX.XX.XX.XX myserver loghost What is the purpose of loghost? If it was not for having loghost in there, all the /etc/hosts files on all the servers in this particular network could be identical. Edit: I looked at /etc/syslog.conf #ident "@(#)syslog.conf 1.5 98/12/14 SMI" /* SunOS 5.0 */ # # Copyright (c) 1991-1998 by Sun Microsystems, Inc. # All rights reserved. # # syslog configuration file. # # This file is processed by m4 so be careful to quote (`') names # that match m4 reserved words. Also, within ifdef's, arguments # containing commas must be quoted. # *.err;kern.notice;auth.notice /dev/sysmsg *.err;kern.debug;daemon.notice;mail.crit /var/adm/messages *.alert;kern.err;daemon.err operator *.alert root *.emerg * # if a non-loghost machine chooses to have authentication messages # sent to the loghost machine, un-comment out the following line: #auth.notice ifdef(`LOGHOST', /var/log/authlog, @loghost) mail.debug ifdef(`LOGHOST', /var/log/syslog, @loghost) # # non-loghost machines will use the following lines to cause "user" # log messages to be logged locally. # ifdef(`LOGHOST', , user.err /dev/sysmsg user.err /var/adm/messages user.alert `root, operator' user.emerg * ) Very interesting. when shutting down,, alerts go to all users probably through *.emerg * Looking at ifdef, it seems that the first parameter checks to see if current machine is a loghost, second parameter is what to do if it is and third parameter is what to do if it is not. Edit: If you want to test a logging rule you can use svcadm restart system-log to restart the logging service and then logger -p notice "test" to send a test log message where notice can be replaced with any type such as user.err, auth.notice, etc.

    Read the article

  • Are ZFS snapshots + S3 a viable backup system for several VMs and general fileserver storage?

    - by AllanA
    I've been tasked with setting up a backup system for my small office (around 12 people). Most of our production stuff is on the AWS cloud, so what I need to back up are some small office/development files (under 100G right now), plus our operational VMs and development, which round out to a bit under 1T. I just need something reliable, convenient, and straightforward. I'm comfortable with Linux, FreeBSD, and to some extent Solaris 10, so I'm leaning toward a full server rather than an appliance system ala Openfiler or FreeNAS. What I'm contemplating is a small fileserver for general storage and nightly backups of the virtual machines, followed up by an offsite backup to Amazon's S3 storage service. It'd be the usual incremental backups nightly and full backup weekly. My question is if using ZFS snapshots, both locally and dumped to S3 via 'zfs send [-i]', is a viable backup tool? Or should I stick to using Duplicity, or some other method entirely? ZFS snapshots on the internal fileserver/backup machine sound like a perfect way to provide quick and convenient data recovery, so I'm likely to go with that for local redundancy. (If you folks see scenarios where relying on ZFS snapshots would be worse than a more traditional archiving backup, feel free to convince me.) But are snapshots flexible enough to lean on for recovery from the loss of my backup server? Or am I better off with something more traditional? (feel free to recommend free or commercial backup solutions you favor.)

    Read the article

  • Command or tool to display list of connections to a Windows file share

    - by BizTalkMama
    Is there a Windows command or tool that can tell me what users or computers are connected to a Windows fileshare? Here's why I'm looking for this: I've run into issues in the past where our deployment team has deployed BizTalk applications to one of our environments using the wrong bindings, leaving us with two receive locations pointing to the same file share (i.e. both dev and test servers point to dev receive location uri). When this occurs, the two environments in question tend to take turns processing the files received (meaning if I am attempting to debug something in one environment and the other environment has picked the file up, it looks as if my test file has disappeared into thin air). We have several different environments, plus individual developer machines, and I'd rather not have to check each individually to find the culprit. I'm looking for a quick way to detect what locations are connected to the share once I notice my test files vanishing. If I can determine the connections that are invalid, I can go directly to the person responsible for that environment and avoid the time it takes to randomly ask around. Or if the connections appear to be correct, I can go directly to troubleshooting where in the process the message gets lost. Any suggestions?

    Read the article

< Previous Page | 430 431 432 433 434 435 436 437 438 439 440 441  | Next Page >