Search Results

Search found 9715 results on 389 pages for 'servers'.

Page 360/389 | < Previous Page | 356 357 358 359 360 361 362 363 364 365 366 367  | Next Page >

  • Intermittent internet access on a flat network - Router is connected

    - by Naveed
    I’m looking for some help with network settings. I’ve just started a new job (non-IT!) and we have problems with our office network. I’m the most IT literate in the organisation (15 permanent employees) and so have been dealing with IT issues. Our main bit of software is web-based so we need constant web access but it sometimes goes down for between 20 minutes and 3 hours despite everything seemingly working fine. It’s a flat network with wireless APs, BT Business Broadband 8Mbit connection and that’s about it. We have no servers and no standard settings and staff are encouraged to bring in their own laptops and connect! The network basically exists to provide internet access and that’s it. We also have students accessing the wireless (and I know there’s a whole list of access and content issues etc but right now we just need internet access stabilised). This is what we have: Building 1 Cisco SLM-224P 24-port PoE 10/100 switch with 2 gigabit ports 3 x ZyXEL NWA-3160 wireless APs Samsung OfficeServ 7100 phone server which borrows the building’s wiring Building 2 Netgear GS605-UK 5-port 10/100/1000 switch 1 x ZyXEL NWA-3160 wireless AP 1 x BT Business Hub – 2wire BT2700hgv – is the DHCP server We have 2 link cables between the buildings. One connects the two switches on a gigabit port. The second (oddly) connects the switch in building 2 to the OfficeServ server in building 1. When the internet goes down I can still access the router through a wireless connection. I can also ping websites and get a response. Firefox just says “Cannot connect” etc. The system then heals itself when it feels like it. (Sorry if this is asking too much but) These are my immediate questions… Why would browser-based internet go down? I don’t know enough about protocols etc but I can try to standardise settings. The WAPs have a DNS server setting and I don’t know whether it should be “None” or “From DHCP”. What should be the DHCP server? The router or the Cisco switch? Or something else?! Would there be any problem in connecting the second link from switch to switch? Is that good practice? Is it worth swapping the Netgear GS605 with either a Cisco SG200-08 or Netgear GS108T-200? Is it worth upgrading the router to, for instance, a Cisco RV042G Dual Gigabit router which would also act as a switch? Or is it better to have a separate router and switch in Building 2?

    Read the article

  • iftop Shows Lots of Mysterious Connections - Not Showing in netstat

    - by HOLOGRAPHICpizza
    I've just stopped all pretty much all services except sshd on my server (Ubuntu Server 10.04), and when I run iftop I get output that looks like this: 12.5Kb 25.0Kb 37.5Kb 50.0Kb 62.5Kb mqqqqqqqqqqqqqqqqvqqqqqqqqqqqqqqqqqvqqqqqqqqqqqqqqqqvqqqqqqqqqqqqqqqqqvqqqqqqqqqqqqqqqqq flash.gateway.2wire.net:ssh <=> 172.16.1.151:60405 1.75Kb 1.54Kb 2.22Kb flash.gateway.2wire.net:21095 <=> 69.127.29.20:32582 536b 107b 27b flash.gateway.2wire.net:21095 <=> 190.164.122.134:13557 0b 105b 26b flash.gateway.2wire.net:21095 <=> 79.165.212.195:45138 0b 105b 26b flash.gateway.2wire.net:21095 <=> 151.42.15.151:9031 0b 72b 18b flash.gateway.2wire.net:21095 <=> 88.185.120.179:51413 0b 0b 49b flash.gateway.2wire.net:21095 <=> 178.120.152.97:25924 0b 0b 29b flash.gateway.2wire.net:21095 <=> 109.110.217.77:27868 0b 0b 26b flash.gateway.2wire.net:21095 <=> 84.13.201.90:16509 0b 0b 26b flash.gateway.2wire.net:21095 <=> 171.7.125.224:11777 0b 0b 26b flash.gateway.2wire.net:21095 <=> 115.177.164.170:21360 0b 0b 26b flash.gateway.2wire.net:21095 <=> 50.88.126.18:25540 0b 0b 25b flash.gateway.2wire.net:21095 <=> 223.206.230.163:13431 0b 0b 25b flash.gateway.2wire.net:21095 <=> 78.144.187.26:24515 0b 0b 25b flash.gateway.2wire.net:21095 <=> 83.20.61.211:27572 0b 0b 25b flash.gateway.2wire.net:21095 <=> 82.134.151.42:18448 0b 0b 18b flash.gateway.2wire.net:21095 <=> 126.117.95.247:25316 0b 0b 18b flash.gateway.2wire.net:21095 <=> 116.202.65.230:9044 0b 0b 18b flash.gateway.2wire.net:21095 <=> 88.120.63.205:51413 0b 0b 17b qqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqq TX: cumm: 61.6KB peak: 8.00Kb rates: 1.59Kb 1.38Kb 2.04Kb RX: 18.4KB 1.64Kb 696b 549b 640b TOTAL: 80.0KB 9.64Kb 2.27Kb 1.92Kb 2.66Kb This is the first part (not the unix socket part) of the output of netstat -a: Active Internet connections (servers and established) Proto Recv-Q Send-Q Local Address Foreign Address State tcp 0 0 *:ssh *:* LISTEN tcp 0 0 *:55677 *:* LISTEN tcp 0 0 flash.gateway.2wire:ssh 172.16.1.151:60405 ESTABLISHED tcp 0 48 flash.gateway.2wire:ssh 172.16.1.151:60661 ESTABLISHED tcp6 0 0 [::]:ssh [::]:* LISTEN udp 0 0 *:37790 *:* What could all those strange connections on port 21095 be? And why would they not show up in netstat?? Any advice would be greatly appreciated.

    Read the article

  • What are some of the best wireless routers for a price-conscious home power-user?

    - by Alain
    I'm extremely dissatisfied with the 'popular' choice for routers in homes and small offices. They are expensive (upwards of 60$), lack a great deal of useful configuration options, and seem to need to be restarted quite often. (Linksys comes to mind). I've been on the market for a good router lately, and slowly collecting a set of requirements I feel good routers should meet. Maximum number of TCP/IP connections. - This isn't something I see any routers advertise, but in terms of supporting torrent applications, I've been screwed by routers that support less than 20 here. From what I understand a fairly standard number is 200, but there are not so expensive routers that support thousands. Router configuration menu - Most have standard menu's that let you set up basic things like your wireless network encryption settings, uPnP, and maybe even DMZ (demilitarized zones). An absolute requirement for me, however, are routers with good enough firmware to support: Explicit Port forwarding Assigning static local ips to specific mac addresses, or at least Port forwarding by MAC address Port, IP and MAC filtering Dynamic DNS service for home users who want to set up a server but have a dynamic IP Traffic shaping (ideally) - giving priority to packets from certain machines or over certain ports. Strong wireless signal - If getting a reliable signal requires me to be so close to the router that I can connect an Ethernet cable, it's not good enough. As many Ethernet ports as possible. - Because I want to be able to switch from console gaming to PC gaming without visiting my router. So far, the best thing I've stumbled upon (in the bargain bin at staples) was a 20$ retail plus router. It was meant to be the cheapest alternative until I could find something better to purchase online, but I was actually blown away by the firmware capabilities. It supports defining reserved bandwidth for certain network traffic, dynamic DNS, reserving local IPs for specific MAC addresses, etc. At 2 am when my roommate is killing our Internet with their torrents, I can limit their bandwidth without outright blacklisting them. I have, however, met serious limitations when it comes to network traffic between local machines. It claims a 300Mbps connection, but I have trouble streaming videos from my PC to my console or other laptops wirelessly. It has a meltdown and needs to be reset once in a while (no more than a couple times a month), and it's got a 200 connection limit. There 4 Ethernet ports in the back but I'm pretty sure the first doesn't work. So some great answers to this question would be: Any metrics you use to compare routers, and requirements you have for new candidates. The best routers you've found for supporting home servers, file management systems, high volume torrent traffic, good price/feature ratio, etc. Good configuration advice (aside from 'use Ethernet whenever possible') Thanks for your feedback and experiences!

    Read the article

  • DNS and DHCP dies after ~2 days of use on ClearOS

    - by TheLQ
    I'm using ClearOS (based on CentOS, so any info specific to it should apply here) as a gateway, DHCP, and DNS server. I had this server running perfectly for a month or two before replacing it with another server. However due DNS and DHCP failing 2 days in and a host of other performance issues (the box was a little underpowered), I changed back to the origional server. However 2 days in DHCP and DNS are failing again, and I'm out of idea's on why. In both cases to my knowledge no network or server changes occurred after installation. Right after installing (and at least a day in) DNS and DHCP was working just fine. However later (Day 2) I get a call saying their internet is down (translation: Nobody can get to websites because DNS is down) I've tried to fix the problem by checking if the dnsmasq is even running (it is), restarting the service, and restarting the server to no effect. I do have two internal servers that have static DHCP leases but one's lease must of expired as I can't connect to it anymore. I'm hesitant to do any dhcp testing on the last server as I'll not be able to connect to it anymore. Is there anything anyone can think of on why DNS and DHCP would fail 2 days in to running perfectly? More info: Running dnsmasq in debug mode. This is all that's displayed even when running nslookup quackwall. I'm not sure though if nslookup commands should show up in the log [root@quackwall ~]# /usr/sbin/dnsmasq -dq dnsmasq: started, version 2.49 cachesize 150 dnsmasq: compile time options: IPv6 GNU-getopt no-DBus no-I18N DHCP TFTP dnsmasq-dhcp: DHCP, IP range 10.0.0.100 -- 10.0.0.254, lease time 12h dnsmasq: reading /etc/resolv.conf dnsmasq: using nameserver 74.128.17.114#53 dnsmasq: using nameserver 74.128.19.102#53 dnsmasq: read /etc/hosts - 5 addresses dnsmasq-dhcp: read /etc/ethers - 2 addresses On the other server DNS and the Gateway are all configured correctly (10.0.0.2 is quackwall) lordquackstar@quackgame:~$ netstat -rn Kernel IP routing table Destination Gateway Genmask Flags MSS Window irtt Iface 10.0.0.0 0.0.0.0 255.255.240.0 U 0 0 0 eth0 0.0.0.0 10.0.0.2 0.0.0.0 UG 0 0 0 eth0 lordquackstar@quackgame:~$ cat /etc/resolv.conf nameserver 10.0.0.2 domain highwow.lan search highwow.lan

    Read the article

  • Installing OpenLDAP on Fedora 12: ldap_bind: Invalid credentials (49)

    - by Alpha Hydrae
    I've been trying to set up the OpenLDAP installed by default on Fedora 12, very unsuccessfully. My ultimate goal is to use LDAP authentication for user login and Apache, using the OpenLDAP server running on the same machine. The server is running, but the error I always get when I try to use ldapsearch or ldapadd is: ldap_bind: Invalid credentials (49) I've been following these tutorials, but none of them helped me: http://www.howtoforge.com/openldap_fedora7 http://www.redhat.com/docs/manuals/linux/RHL-9-Manual/ref-guide/s1-ldap-quickstart.html http://www.howtoforge.com/linux_ldap_authentication http://docs.fedoraproject.org/deployment-guide/f12/en-US/html/s1-ldap-pam.html http://www.openldap.org/doc/admin24/quickstart.html First, some components were already installed, and I installed these with yum: yum install openldap-servers openldap-devel Then, I created a basic slapd.conf file in /etc/openldap: database bdb suffix "dc=sniejana-sandbox,dc=com" rootdn "cn=root,dc=sniejana-sandbox,dc=com" rootpw {SSHA}cxdz55ygPu4T3ykg7dgu+L0VRvsFSeom directory /var/lib/ldap/sniejana-sandbox.com I obtained the rootpw with this command: slappasswd -s changeme I also created the /var/lib/ldap/sniejana-sandbox.com directory and made sure the entire contents of /var/lib/ldap were owned by the ldap user. I found two ldap.conf files, one in /etc and one in /etc/openldap. I don't know which is the right one. If I understood correctly, this file is to configure the client. I put this in both: HOST localhost BASE dc=sniejana-sandbox,dc=com I then ran the server with: service slapd start It said OK. Most of the tutorials above say to use the command ldapsearch -D "cn=Manager,dc=my-domain,dc=com" -W to ensure that everything's working. When I execute this command, a password prompt appears, and after entering the password, I get the error. ldapsearch -D "cn=root,dc=sniejana-sandbox,dc=com" -W Enter LDAP password: ldap_bind: Invalid credentials (49) The same thing happens when trying to use ldapadd. I tried with an encrypted and unencrypted password in slapd.conf, it doesn't change anything. Adding a -x for simple authentication doesn't change anything either. netstat -ap confirms the server is listening: tcp 0 0 *:ldap *:* LISTEN 4148/slapd tcp 0 0 *:ldap *:* LISTEN 4148/slapd ps -ef|grep slapd confirms the process is running: ldap 4148 1 0 15:22 ? 00:00:00 /usr/sbin/slapd -h ldap:/// -u ldap Running slaptest procudes config file testing succeeded. I read somewhere that the command ldapsearch -x -b '' -s base '(objectclass=*)' namingContext can confirm the server is running. It appears to work: # extended LDIF # # LDAPv3 # base <> with scope baseObject # filter: (objectclass=*) # requesting: namingContext # # dn: # search result search: 2 result: 0 Success # numResponses: 2 # numEntries: 1 I'm running out of ideas. Am I missing something obvious?

    Read the article

  • Apache/Jboss Issue - is this connection timeout?

    - by user115391
    We have an application. The architecture is as below 1 load balancer (apache), which redirects to 2 app servers (jboss). The site is working fine and I am able to access it fine. But sometimes, randomly the homepage takes a while (like 30-40 secs) to load. I tried checking the logs but could not figure out why. I used the httptraffic analyzer, fiddler to see the traffic, but it just says the request/response took 30 secs or so. I checked the apache access logs, mod_jk.log. My configurations are below mod-jk.conf LoadModule jk_module modules/mod_jk.so JkWorkersFile conf/workers.properties JkLogFile logs/mod_jk.log #JkLogLevel info #JkLogLevel debug JkLogLevel error # Select the log format JkLogStampFormat "[%a %b %d %H:%M:%S %Y]" JkOptions +ForwardKeySize +ForwardURICompatUnparsed -ForwardDirectories JkRequestLogFormat "%w %V %T %P %{tid}P %D" JkMount /__application__/* loadbalancer JkUnMount /__application__/images/* loadbalancer <VirtualHost *:8080 > JkMountFile conf/uriworkermap.properties </VirtualHost> JkShmFile run/jk.shm <Location /jkstatus> JkMount status Order deny,allow Deny from all Allow from 127.0.0.1 </Location> ----------------------------- uriworkermap.properties Simple worker configuration file # Mount the Servlet context to the ajp13 worker /=loadbalancer /*=loadbalancer ----------------------------- workers.properties worker.list=loadbalancer,status worker.template.port=8009 worker.template.type=ajp13 worker.template.lbfactor=1 worker.template.prepost_timeout=10000 worker.template.connect_timeout=10000 worker.template.ping_mode=A worker.worker1.reference=worker.template worker.worker1.host=hostname1 worker.worker2.reference=worker.template worker.worker2.host=hostname2 worker.loadbalancer.type=lb worker.loadbalancer.balance_workers=worker1,worker2 worker.status.type=status ----------------------------- my jboss server.xml - $JBOSS_HOME/server/default/deploy/jbossweb.sar/server.xml --------------------------------- The logs from access log is below The issue where it took time - look at the seconds column [23/Mar/2012:12:10:38 -0400] "GET / HTTP/1.1" 200 138 x.x.x.x - - [23/Mar/2012:12:10:49 -0400] "GET /index.jsp HTTP/1.1" 302 - x.x.x.x - - [23/Mar/2012:12:11:10 -0400] "GET /home.jsp HTTP/1.1" 200 936 x.x.x.x - - [23/Mar/2012:12:11:31 -0400] "POST /login/ HTTP/1.1" 200 8895 x.x.x.x - - [23/Mar/2012:12:11:52 -0400] "GET /login/includes/login-style.css HTTP/1.1" 304 - The one after the issue x.x.x.x - - [23/Mar/2012:12:12:18 -0400] "GET / HTTP/1.1" 200 138 x.x.x.x - - [23/Mar/2012:12:12:18 -0400] "GET /index.jsp HTTP/1.1" 302 - x.x.x.x - - [23/Mar/2012:12:12:18 -0400] "GET /home.jsp HTTP/1.1" 200 936 x.x.x.x - - [23/Mar/2012:12:12:18 -0400] "POST /login/ HTTP/1.1" 200 8895 x.x.x.x - - [23/Mar/2012:12:12:18 -0400] "GET /login/includes/login-style.css HTTP/1.1" 304 - Would it be a cache or timeout issue? Any help is appreciated. Thanks.

    Read the article

  • Ubuntu Server 10.04 Heavy Network Traffic causes disconnect

    - by K Vaughan
    I'm currently running a headless Ubuntu 10.04 server. Installed is the LAMP stack, Joomla, Virtualbox, phpvirtualbox, webmin and proFTP.. It resolves the IP address so I can access it remotely (either the apache2 webserver or the FTP) using DDClient. Any packages installed have been installed using apt-get. Webmin, although discouraged in Ubuntu Server, is used mostly to administer the webserver aspect. This issue also appeared when I was using Ubuntu Server 10.10. After periods of heavy network traffic, whether local or remote, the connect drops. I'm talking specifically about the transfer of files via FTP, SCP or Samba (the latter of which I seldom use). There is no response to ping or ssh. I can't FTP to the server nor can I load the website. There are times when the server has been on for a few days and everything runs fine because I haven't accessed it much, if at all (thus not much network traffic). I've gone through a few hardware changes although I don't believe this has cause the issue: this has been happening long before I made any changes. At first I thought it was my ISP-provided router blocking traffic because of some kind of misconfiguration (perhaps assuming it was some kind of DoS attack). I've changed routers and still found no success. I've checked syslog, dmesg and kern.log for warnings but have uncovered none. I've ran memtest via the GRUB2 menu at boot and once it turned up 4 errors. I ran again with individual sticks of RAM in various slots and everything turned up fine. I've looked through the BIOS settings and everything looks fine. I've tried unplugging unnecessary pieces of hardware (other internal hard drives, CD drives, floppy, PCI cards, etc). Any help or tips on how I can even begin to troubleshoot this would be very much appreciated. Please note that i've only started playing with servers as a hobby so my knowledge wouldn't be the most refined. I'm comfortable with command line and have the initiative to know how to look up something I can't do. Unfortunately I can't seem to find any issues like this. Additionally: If a solution can't be found some assistance to write a script that will cause the server to reboot automatically if, after x minutes, it gets no response to pinging somewhere like google. Admittedly that's not the cleanest solution should my internet end up going down but I can't think of what else to do.

    Read the article

  • How to have Windows Server DNS use hosts file to resolve specific host names

    - by user41079
    Hello, everyone, I'm facing a small problem with Windows Server 2003 DNS service. In my corporation, I'm running Microsoft DNS server(172.16.0.12) to do name resolution to my company intranet(domain name ends in dev.nls. resolving to IP 172.16..), and it is also configured as a DNS forwarder to forward other domain names(e.g. *.google.com , *.sf.net) to Internet real DNS servers. This internal DNS server never tends to serve users from outside world. And, we are running a mail server(serving incoming mail for a real Internet domain @nlscan.com) inside company firewall which can be accessed in either way: by connecting to 172.16.0.10 from within intranet. by connecting to mail.nlscan.com(resolved to 202.101.116.9) from Internet. Note that 172.16.0.10 and 202.101.116.9 is not the same physical machine. The 202 one is a firewall machine who do port forwarding of port 25 and 110 to intranet address 172.16.0.10 . Now my question: If users inside corporate LAN want to resolve mail.nlscan.com, it resolves to 202.101.116.9. That's correct and workable, BUT NOT GOOD, because the mail traffic goes to the firewall machine then bounces to 172.16.0.10 . I hope that our internal DNS server can intercept the name mail.nlscan.com and resolve it to 172.16.0.10 . So, I hope that I can write an entry in "hosts" file on 172.16.0.12 to do this. But, how can Microsoft DNS server recognize this "hosts" file? Maybe you suggest, why not have intranet user use 172.16.0.10 to access my mail server? I have to say it is inconvenient, suppose a user(employee) works on his laptop, daytime in office and nighttime at home. When he is at home, he cannot use 172.16.0.10 . Creating a zone for nlscan.com on our internal DNS server is not feasible, because the name server for nlscan.com domain is on our ISP, and it is responsible for resolving other host names and sub-domains under nlscan.com . Thank you in advance.

    Read the article

  • Change default DNS server in Arch Linux

    - by AntoineG
    I'm in Viet Nam and most social websites (Facebook, Twitter and the likes - even reddit) are blocked by the ISP DNS server. I tried to change the DNS server of my Arch box using the resolv.conf file, but it failed miserably since dhcpd generates this file automatically everytime I connect to the LAN. I've been looking around to try and find out how to fix this, without success. Either I s*ck at Googling, either it is non-trivial to do so. EDIT 1: Meh, apparently posting it here made me feel guilty and I had to push my search a bit more. I found the same article than Ankur post below. This is what I made, if anybody ever faces the same problem: $ sudo gvim /etc/dhcpcd.conf Add "nohook resolv.conf" at the tail of the file. $ sudo gvim /etc/resolv.conf Add to the file (OpenDNS servers): nameserver 208.67.222.222 nameserver 208.67.220.220 Or (Google DNS): nameserver 8.8.8.8 nameserver 8.8.4.4 Then, verify it worked (need package dnsutils): $ dig www.facebook.com ; <<>> DiG 9.9.1-P1 <<>> www.facebook.com ;; global options: +cmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 16994 ;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1 ;; OPT PSEUDOSECTION: ; EDNS: version: 0, flags:; udp: 4096 ;; QUESTION SECTION: ;www.facebook.com. IN A ;; ANSWER SECTION: www.facebook.com. 89 IN A 69.171.224.53 ;; Query time: 87 msec ;; SERVER: 208.67.222.222#53(208.67.222.222) ;; WHEN: Thu Jun 28 00:43:23 2012 ;; MSG SIZE rcvd: 61 See ;; SERVER: 208.67.222.222#53(208.67.222.222), it worked.

    Read the article

  • Slow Memcached: Average 10ms memcached `get`

    - by Chris W.
    We're using Newrelic to measure our Python/Django application performance. Newrelic is reporting that across our system "Memcached" is taking an average of 12ms to respond to commands. Drilling down into the top dozen or so web views (by # of requests) I can see that some Memcache get take up to 30ms; I can't find a single use of Memcache get that returns in less than 10ms. More details on the system architecture: Currently we have four application servers each of which has a memcached member. All four memcached members participate in a memcache cluster. We're running on a cloud hosting provider and all traffic is running across the "internal" network (via "internal" IPs) When I ping from one application server to another the responses are in ~0.5ms Isn't 10ms a slow response time for Memcached? As far as I understand if you think "Memcache is too slow" then "you're doing it wrong". So am I doing it wrong? Here's the output of the memcache-top command: memcache-top v0.7 (default port: 11211, color: on, refresh: 3 seconds) INSTANCE USAGE HIT % CONN TIME EVICT/s GETS/s SETS/s READ/s WRITE/s cache1:11211 37.1% 62.7% 10 5.3ms 0.0 73 9 3958 84.6K cache2:11211 42.4% 60.8% 11 4.4ms 0.0 46 12 3848 62.2K cache3:11211 37.5% 66.5% 12 4.2ms 0.0 75 17 6056 170.4K AVERAGE: 39.0% 63.3% 11 4.6ms 0.0 64 13 4620 105.7K TOTAL: 0.1GB/ 0.4GB 33 13.9ms 0.0 193 38 13.5K 317.2K (ctrl-c to quit.) ** Here is the output of the top command on one machine: ** (Roughly the same on all cluster machines. As you can see there is very low CPU utilization, because these machines only run memcache.) top - 21:48:56 up 1 day, 4:56, 1 user, load average: 0.01, 0.06, 0.05 Tasks: 70 total, 1 running, 69 sleeping, 0 stopped, 0 zombie Cpu(s): 0.0%us, 0.0%sy, 0.0%ni, 99.7%id, 0.0%wa, 0.0%hi, 0.0%si, 0.3%st Mem: 501392k total, 424940k used, 76452k free, 66416k buffers Swap: 499996k total, 13064k used, 486932k free, 181168k cached PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 6519 nobody 20 0 384m 74m 880 S 1.0 15.3 18:22.97 memcached 3 root 20 0 0 0 0 S 0.3 0.0 0:38.03 ksoftirqd/0 1 root 20 0 24332 1552 776 S 0.0 0.3 0:00.56 init 2 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kthreadd 4 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kworker/0:0 5 root 20 0 0 0 0 S 0.0 0.0 0:00.02 kworker/u:0 6 root RT 0 0 0 0 S 0.0 0.0 0:00.00 migration/0 7 root RT 0 0 0 0 S 0.0 0.0 0:00.62 watchdog/0 8 root 0 -20 0 0 0 S 0.0 0.0 0:00.00 cpuset 9 root 0 -20 0 0 0 S 0.0 0.0 0:00.00 khelper ...output truncated...

    Read the article

  • NTPD issue - syncs then slowly loses ground

    - by ethrbunny
    RHEL 5 workstation. Has been running smoothly for years. I did a 'pup' recently and followed with a nice, cleansing reboot. Afterwards the system had some startup issues: namely MySQL refused to start. It just went "...." for 5-10 minutes before I did another boot and skipped that step (using 'interactive'). This was the only service that didn't wan't to start normally. So now that the system is booted I've found that it doesn't want to stay in sync with the NTP master and after 48 hours is refusing any SSH other than root. NTPD: this service starts normally and gets a lock on 4 servers. Almost immediately it starts to lose ground and now (after 3 days) is almost 40 hours behind. If I stop/start the service it gets the lock, resets the system clock and starts losing ground again. The 'hwclock' is set properly and maintains its time. Login: when I (re)start the ntp server I am able to login normally. I assume this problem is due to losing sync with LDAP. This appears to be verified by LDAP errors in /var/log/messages. Suggestions on where to look? ADDENDA: Tried deleting the 'drift' file. After a bit it gets recreated with 0.000. from /var/log/messages: Jan 17 06:54:01 aeolus ntpdate[5084]: step time server 129.95.96.10 offset 30.139216 sec Jan 17 06:54:01 aeolus ntpd[5086]: ntpd [email protected] Tue Oct 25 12:54:17 UTC 2011 (1) Jan 17 06:54:01 aeolus ntpd[5087]: precision = 1.000 usec Jan 17 06:54:01 aeolus ntpd[5087]: Listening on interface wildcard, 0.0.0.0#123 Disabled Jan 17 06:54:01 aeolus ntpd[5087]: Listening on interface wildcard, ::#123 Disabled Jan 17 06:54:01 aeolus ntpd[5087]: Listening on interface lo, ::1#123 Enabled Jan 17 06:54:01 aeolus ntpd[5087]: Listening on interface eth0, fe80::213:72ff:fe20:4080#123 Enabled Jan 17 06:54:01 aeolus ntpd[5087]: Listening on interface lo, 127.0.0.1#123 Enabled Jan 17 06:54:01 aeolus ntpd[5087]: Listening on interface eth0, 10.127.24.81#123 Enabled Jan 17 06:54:01 aeolus ntpd[5087]: kernel time sync status 0040 Jan 17 06:54:02 aeolus ntpd[5087]: frequency initialized 0.000 PPM from /var/lib/ntp/drift Jan 17 06:54:02 aeolus ntpd[5087]: system event 'event_restart' (0x01) status 'sync_alarm, sync_unspec, 1 event, event_unspec' (0xc010) You can see the 30 second offset. This was after about one minute of operation.

    Read the article

  • Windows XP long login (15 minutes +)

    - by Emily Pinkerton
    I'm having a lot of issues with our Windows XP SP3 machines (about 5, but every week another gets on the bandwagon of this issue). They take forever (15 minutes) to apply the user settings once our employee's enter their username and password to login to our domain. It only happens say if a user has reboot the machine and then when they go to log back in then it hangs forever. Reboot and restart are the key words for sure I've noticed with this issue. Here are things I have tested: •Made sure the DNS was set to point to our two servers (Server01 & Server02 are DNS Domain Controllers, 01 is primary and 02 backup). •No major changes have been applied to our network. •All profiles are local, so I have deleted out local profiles that aren't being used on those machines that run slow. •Also I have tried to enable and disable the Enable Fast Login under the local machines GP. It was not configured originally and when I tested both, it made the computer hang on "applying computer settings" for about 15 minutes. When it finally came up to the login screen the it was very quick to login to the domain. However this doesn't fix my issue, and even more frustrating upon setting it back to being not configured it now still takes for forever to apply computer settings. •I enabled the userenv log and here is what I see, but my experience is limited and I'm not sure how to read it exactly. (see below for log, this isn't the whole thing because it's really long) USERENV(2ec.2f0) 10:50:41:843 LoadUserProfile: LoadUserProfileP succeeded USERENV(2ec.2f0) 10:50:41:843 LoadUserProfile: Returning success. Final Information follows: USERENV(2ec.2f0) 10:50:41:843 lpProfileInfo-UserName = USERENV(2ec.2f0) 10:50:41:843 lpProfileInfo-lpProfilePath = < USERENV(2ec.2f0) 10:50:41:843 lpProfileInfo-dwFlags = 0x0 USERENV(2ec.2f0) 10:50:41:843 LoadUserProfile: Returning TRUE. hProfile = <0x818 USERENV(2ec.2f0) 10:50:41:984 IsSyncForegroundPolicyRefresh: Synchronous, Reason:NonCachedCredentials USERENV(2ec.248) 10:50:41:984 IsSyncForegroundPolicyRefresh: Synchronous, Reason:NonCachedCredentials USERENV(3c4.3dc) 10:51:26:166 LibMain: Process Name: C:\WINDOWS\system\wbem\wmiprvse.exe USERENV(2ec.5cc) 11:05:08:741 ProcessGPOs: network name is 192.168.49.0 USERENV(4a8.888) 11:05:08:804 GetProfileType: Profile already loaded. USERENV(4a8.888) 11:05:08:804 LoadProfileInfo: Failed to query central profile with error 2 USERENV(4a8.888) 11:05:08:804 GetProfileType: ProfileFlags is 0 Also this error is in the file quite a lot: USERENV(328.5bc) 11:05:29:733 GetUserDNSDomainName: Failed to impersonate user USERENV(328.834) 11:05:29:733 ImpersonateUser: Failed to impersonate user with 5. I'm really not sure what else to do with my limited experience, but I'm hoping someone can help me. I feel like I'm dealing with an issue way above my level and any knowledge I can gain out of getting this issue fixed would be amazing.

    Read the article

  • DHCP server with multiple interfaces on ubuntu, destroys default gateway

    - by Henrik Kjus Alstad
    I use Ubuntu, and I have many interfaces. eth0, which is my internet connection, and it gets its info from a DHCP-server totally outisde of my control. I then have eth1,eth2,eth3 and eth4 which I have created a DHCP-server for.(ISC DHCP-Server) It seems to work, and I even get an IP-address from the foreign DHCP-server on the internet facing interface. However, for some reason it seems my gateway for eth0 became screwed after I installed my local DHCP-server for eth1-eth4. (I think so because I got an IP for eth0, and I can ping other stuff on the local network, but I cannot get access to the internet). My eth0-specific info in /etc/network/interfaces: auto lo iface lo inet loopback auto eth0 iface eth0 inet dhcp auto eth1 iface eth1 inet static address 10.0.1.1 netmask 255.255.255.0 network 10.0.1.0 broadcast 10.0.1.255 gateway 10.0.1.1 mtu 8192 auto eth2 iface eth2 inet static address 10.0.2.1 netmask 255.255.255.0 network 10.0.2.0 broadcast 10.0.2.255 gateway 10.0.2.1 mtu 8192 My /etc/default/isc-dhcp-server: INTERFACES="eth1 eth2 eth3 eth4" So why does my local DHCP-server fuck up the gateway for eth0, when I tell it not to listen to eth0? Anyone see the problem or what I can do to fix it? The problem seems indeed to be the gateways. "netstat -nr" gives: 0.0.0.0 --- 10.X.X.X ---- 0.0.0.0 --- UG 0 0 0 eth3 It should have been 0.0.0.0 129.2XX.X.X 0.0.0.0 UG 0 0 0 eth0 So for some reason, my local DHCP-server overrides the gateway I get from the network DHCP. Edit: dhcp.conf looks like this(I included info only for eth1 subnet): ddns-update-style none; not authoritative; subnet 10.0.1.0 netmask 255.255.255.0 { interface eth1; option domain-name "example.org"; option domain-name-servers ns1.example.org, ns2.example.org; default-lease-time 600; max-lease-time 7200; range 10.0.1.10 10.0.1.100; host camera1_1 { hardware ethernet 00:30:53:11:24:6E; fixed-address 10.0.1.10; } host camera2_1 { hardware ethernet 00:30:53:10:16:70; fixed-address 10.0.1.11; } } Also, it seems that the gateway is correctly set if I run "/etc/init.d/networking restart" in a terminal, but that's not helpful for me, I need the correct gateway to be set during startup, and i'd rather find the source of the problem

    Read the article

  • Unable to connect to SVN server on VPS : Subversion Configuration Problem

    - by Pritam Barhate
    Hello everybody, I purchased a VPS account (centos-5-x86_64) from Hostgator mainly for the purpose of setting up an online Subversion Server. This is the first time I am managing a VPS and my linux skills are not so great, but I have basic knowledge of the OS and have been using it on and off as desktop from last few years. So after foxing through a few tutorials online, as the first step I logged in using SSH root account provided by Hostgator and tried to run, yum install mod_dav_svn subversion As it turns out my account has Cpanel/WHM and since it some concept of easy apache straight forward procedure of yum install mod_dav_svn subversion won't work. After that I found out how it can be worked out by compiling the source and stuff. But the whole procedure looked long and scary. [I am just a linux nub]. so I decided I would just skip whole apache integration stuff and just access the server using svn:// protocol, anyways that's how I configure svn on our LAN. So I installed subversion using yum install subversion It installed fine. Then I created a folder /svn_repos/testproject and ran svnadmin create /svn_repos/testproject/ No Problems Using vi I changed svnserve.conf and passwd files for the repository and added a user with my name. Anonymous users don't have any access, authenticated users have write access. Then I started svnserve using svnserve -d then in same terminal window svn list svn://localhost/svn_repos/testproject Asks for authentication for realm, provided root password then for svn username and password. Provided both. The command returns nothing but exists properly. Returns nothing is understood I didn't import anything. But if try to access svn remotely using in another terminal: svn list svn://ip.add.of.server/svn_repos/testproject svn: Can't connect to host 'ip.add.of.server': Operation timed out Is what I get. Parallels Power Panel that I got from Hostgator reports that: The firewall is not active now. To activate the firewall, choose one of the firewall operation modes. So if firewall is not running and I can access svn using localhost, why the operation is timing out when I try to access svn from a remote machine? Experienced network admins please help. Thanks in advance. Also please suggest a good book which gives detailed information on configuring Dedicated servers + WHM and CPanel.

    Read the article

  • Can't ping devices by IP address for devices allocated IPs by DHCP

    - by GiddyUpHorsey
    I have a home network with a Trendnet wireless router and a Windows Domain. The Domain Controller/DNS server is a Windows 2000 Server and is configured to forward queries to DNS servers provided by the ISP. The router provides DHCP and is configured with the Windows 2000 Server as the DNS server. The network has been set up for a couple of years and usually works fine. When I connect iPhones to the network over WiFi, the router can ping the iPhones through its browser based admin interface, but Windows machines that are part of the Windows Domain cannot. A laptop was connected to the network over WiFi that wasn't joined to the domain and it could see the iPhones. The router UI shows that the laptop has a reserved IP allocated via DHCP. All machines either have a static or DHCP allocated IP on the 192.168.0.* subnet. Router - 192.168.0.1 - Static - Wired Windows Domain Controller - 192.168.0.8 - Static - Virtual Windows 7 Workstation - 192.168.0.200 - DHCP Auto - Wired VMWare ESXi Host - 192.168.0.201 - Static? - Wired iPhone 1 - 192.168.0.202 - DHCP Auto - WiFi iPhone 2 - 192.168.0.203 - DHCP Auto - WiFi Windows Vista Laptop - 192.168.0.204 - DHCP Reserved - WiFi Using the Windows 7 machine (200), I try to ping each machine and the only DHCP machine that responds is itself. The other DHCP machines fail with Reply from 192.168.0.200: Destination host unreachable.. Using nslookup fails with *** domain.controller.name can't find 192.168.0.203: Non-existent domain. Using the Windows 2000 Domain Controller (8), I try to ping each machine and the only DHCP machine that responds is the Windows 7 machine (200). Pinging the other DHCP machines fails with Request timed out.. Using nslookup also fails with *** domain.controller.name can't find 192.168.0.203: Non-existent domain. Using the iPhone 2 (203), I try to ping (Network Ping Lite) the machines with static IP addresses and that works fine. When I try to ping the Windows 7 machine (200) it is unable to get a response. How do I configure the DNS server/Windows Domain/Router properly so that the Windows Domain machines can see the IPs allocated via DHCP?

    Read the article

  • IPvsadm not equally balancing on wlc scheduler

    - by davidsmalley
    For some reason, ipvsadm does not seem to be equally balancing the connections between my real servers when using the wlc or lc schedulers. One real server gets absolutely hammered with requests while the others receive relatively few connections. My ldirectord.cf file looks like this: quiescent = yes autoreload = yes checktimeout = 10 checkinterval = 10 # *.site.com http virtual = 111.111.111.111:http real = 10.10.10.1:http ipip 10 real = 10.10.10.2:http ipip 10 real = 10.10.10.3:http ipip 10 real = 10.10.10.4:http ipip 10 real = 10.10.10.5:http ipip 10 scheduler = lc protocol = tcp service = http checktype = negotiate request = "/lb" receive = "Up and running" virtualhost = "site.com" fallback = 127.0.0.1:http The weird thing that I think may be causing the problem (but I'm really not sure) is that ipvsadm doesn't seem to be tracking active connections properly, they all appear as inactive connections IP Virtual Server version 1.2.1 (size=4096) Prot LocalAddress:Port Scheduler Flags -> RemoteAddress:Port Forward Weight ActiveConn InActConn TCP 111.111.111.111:http lc -> 10.10.10.1:http Tunnel 10 0 10 -> 10.10.10.2:http Tunnel 10 0 18 -> 10.10.10.3:http Tunnel 10 0 3 -> 10.10.10.4:http Tunnel 10 0 10 -> 10.10.10.5:http Tunnel 10 0 5 If I do ipvsadm -Lnc then I see lots of connections but only ever in ESTABLISHED & FIN_WAIT states. I was using ldirectord previously on a Gentoo based load balancer and the activeconn used to be accurate, since moving to Ubuntu 10.4 LTS something seems to be different. # ipvsadm -v ipvsadm v1.25 2008/5/15 (compiled with popt and IPVS v1.2.1) So, is ipvsadm not tracking active connections properly and thus making load balancing work incorrectly and if so, how do I get it to work properly again? Edit: It gets weirder, if I cat /proc/net/ip_vs then it looks like the correct activeconns are there IP Virtual Server version 1.2.1 (size=4096) Prot LocalAddress:Port Scheduler Flags -> RemoteAddress:Port Forward Weight ActiveConn InActConn TCP B86A9732:0050 rr -> 0AB42453:0050 Tunnel 10 1 24 -> 0AB4321D:0050 Tunnel 10 0 23 -> 0AB426B2:0050 Tunnel 10 2 25 -> 0AB4244C:0050 Tunnel 10 2 22 -> 0AB42024:0050 Tunnel 10 2 23

    Read the article

  • Unusual Apache->Tomcat caching issue.

    - by iftrue
    Right now, I have an Apache setup sitting in front of Tomcat to handle caching. This setup has been given to an external service to manage, and since the transition, I've noticed odd behavior. Specifically, when I request a swf file from the web server, I hit the Apache cache (good), but occasionally I'll receive a truncated file. Once I receive this truncated file, the cache will NOT refresh until I manually delete the cache and let the swf pull down from tomcat again. The external service claims that the configuration is fine, but I don't see any way this could be happening aside from improper configuration. Now, there are two apache and two tomcat servers under a load balancer, and occasionally one apache cache will break while another does not (leading to 50% of all requests getting bad, truncated data). Where should I start looking to debug this issue? What could POSSIBLY be causing this odd behavior? Edit: Inspecting the logs, tomcat throws this: java.io.IOException: Bad file number at java.io.FileInputStream.readBytes(Native Method) at java.io.FileInputStream.read(FileInputStream.java:199) at java.io.BufferedInputStream.read1(BufferedInputStream.java:256) at java.io.BufferedInputStream.read(BufferedInputStream.java:317) at java.io.FilterInputStream.read(FilterInputStream.java:90) at org.apache.catalina.servlets.DefaultServlet.copyRange(DefaultServlet.java:1968) at org.apache.catalina.servlets.DefaultServlet.copy(DefaultServlet.java:1714) at org.apache.catalina.servlets.DefaultServlet.serveResource(DefaultServlet.java:809) at org.apache.catalina.servlets.DefaultServlet.doGet(DefaultServlet.java:325) at javax.servlet.http.HttpServlet.service(HttpServlet.java:690) at javax.servlet.http.HttpServlet.service(HttpServlet.java:803) at org.apache.catalina.core.ApplicationFilterChain.internalDoFilter(ApplicationFilterChain.java:290) at org.apache.catalina.core.ApplicationFilterChain.doFilter(ApplicationFilterChain.java:206) at org.apache.catalina.core.StandardWrapperValve.invoke(StandardWrapperValve.java:233) at org.apache.catalina.core.StandardContextValve.invoke(StandardContextValve.java:175) at org.apache.catalina.core.StandardHostValve.invoke(StandardHostValve.java:128) at org.apache.catalina.valves.ErrorReportValve.invoke(ErrorReportValve.java:102) at org.apache.catalina.valves.AccessLogValve.invoke(AccessLogValve.java:568) at org.apache.catalina.core.StandardEngineValve.invoke(StandardEngineValve.java:109) at org.apache.catalina.ha.session.JvmRouteBinderValve.invoke(JvmRouteBinderValve.java:209) at org.apache.catalina.ha.tcp.ReplicationValve.invoke(ReplicationValve.java:347) at org.terracotta.modules.tomcat.tomcat_5_5.SessionValve55.invoke(SessionValve55.java:57) at org.apache.catalina.connector.CoyoteAdapter.service(CoyoteAdapter.java:286) at org.apache.jk.server.JkCoyoteHandler.invoke(JkCoyoteHandler.java:190) at org.apache.jk.common.HandlerRequest.invoke(HandlerRequest.java:283) at org.apache.jk.common.ChannelSocket.invoke(ChannelSocket.java:767) at org.apache.jk.common.ChannelSocket.processConnection(ChannelSocket.java:697) at org.apache.jk.common.ChannelSocket$SocketConnection.runIt(ChannelSocket.java:889) at org.apache.tomcat.util.threads.ThreadPool$ControlRunnable.run(ThreadPool.java:690) at java.lang.Thread.run(Thread.java:619) followed by access_log.2009-12-14.txt:1.2.3.4 - - [14/Dec/2009:00:27:32 -0500] "GET /myApp/mySwf.swf HTTP/1.1" 304 - access_log.2009-12-14.txt:1.2.3.4 - - [14/Dec/2009:01:27:33 -0500] "GET /myApp/mySwf.swf HTTP/1.1" 304 - access_log.2009-12-14.txt:1.2.3.4 - - [14/Dec/2009:01:39:53 -0500] "GET /myApp/mySwf.swf HTTP/1.1" 304 - access_log.2009-12-14.txt:1.2.3.4 - - [14/Dec/2009:02:27:38 -0500] "GET /myApp/mySwf.swf HTTP/1.1" 304 - So apache is caching the bad file size. What could possibly be causing this, and possibly separate, how do I ensure that this exception does not get written to cache?

    Read the article

  • Enterprise Tape Backup solutions

    - by Tom O'Connor
    I'm currently attempting to re-architect a backup solution where I'm working. We've got 2 NAS devices, one in the office, one in the datacentre. The servers in the DC back up to the DC NAS, which is then replicated to the Office NAS. The office NAS exports shares as CIFS and NFS, this bit is fine. At some point, I'll have to expand our storage capacity, currently we've got about 1.4TB of storage space, which is about 96% full. Previously, the tape backup was a script that ran tar a few times and squirted data onto a tape. It worked, but was by no means a perfect solution. Restores are a bit of a pest, adding new data to the backup requires editing the script as root. It's just all a bit non-ideal. I've been evaluating a number of "enterprise" ready backup solutions, such as Yosemite Backup from Barracuda, Acronis Backup/Restore, and something from Arkeia. In the process of evaluating these, I've found 2 big problems. Not all of them allow backup of mounted devices (such as a NFS mounted NAS) Many of these applications don't like our tape device. For the most part, (1) is essential. Our NAS has a feeble processor and can't run applications like backup agents. I suspect that the biggest problem is the tape device, which is a HP C7438A DAT72 connected via USB. Questions: Has anyone else got an USB DAT72 device working with similar software? Is there a better way to back up data from an "appliance" NAS device on which you can't run an agent? Would I be totally out of my mind to specify a cheap HP or Dell server with a couple of 1TB hard disks, and a SAS card to then talk to an HP Ultrium (or similar) device? The biggest drawback to this would be cost (400ish for the server, 200 for the SAS connectivity and 1700 for a LTO4 device) Notes: I'd love to be able to say that I'd get rid of tapes entirely, and use some form of hard disk backup. In a previous job, we had LaCie USB drives, which were decidedly unreliable.

    Read the article

  • open-sshd service withou pam support !! How can I add pam support to sshd? Ubuntu

    - by marc.riera
    Hi, I'm using AD as my user account server with ldap. Most of the servers run with UsePam yes except this one, it has lack of pam support on sshd. root@linserv9:~# ldd /usr/sbin/sshd linux-vdso.so.1 => (0x00007fff621fe000) libutil.so.1 => /lib/libutil.so.1 (0x00007fd759d0b000) libz.so.1 => /usr/lib/libz.so.1 (0x00007fd759af4000) libnsl.so.1 => /lib/libnsl.so.1 (0x00007fd7598db000) libcrypto.so.0.9.8 => /usr/lib/libcrypto.so.0.9.8 (0x00007fd75955b000) libcrypt.so.1 => /lib/libcrypt.so.1 (0x00007fd759323000) libc.so.6 => /lib/libc.so.6 (0x00007fd758fc1000) libdl.so.2 => /lib/libdl.so.2 (0x00007fd758dbd000) /lib64/ld-linux-x86-64.so.2 (0x00007fd759f0e000) I have this packages installed root@linserv9:~# dpkg -l|grep -E 'pam|ssh' ii denyhosts 2.6-2.1 an utility to help sys admins thwart ssh hac ii libpam-modules 0.99.7.1-5ubuntu6.1 Pluggable Authentication Modules for PAM ii libpam-runtime 0.99.7.1-5ubuntu6.1 Runtime support for the PAM library ii libpam-ssh 1.91.0-9.2 enable SSO behavior for ssh and pam ii libpam0g 0.99.7.1-5ubuntu6.1 Pluggable Authentication Modules library ii libpam0g-dev 0.99.7.1-5ubuntu6.1 Development files for PAM ii openssh-blacklist 0.1-1ubuntu0.8.04.1 list of blacklisted OpenSSH RSA and DSA keys ii openssh-client 1:4.7p1-8ubuntu1.2 secure shell client, an rlogin/rsh/rcp repla ii openssh-server 1:4.7p1-8ubuntu1.2 secure shell server, an rshd replacement ii quest-openssh 5.2p1_q13-1 Secure shell root@linserv9:~# What I'm doing wrong? thanks. Edit: root@linserv9:~# cat /etc/pam.d/sshd # PAM configuration for the Secure Shell service # Read environment variables from /etc/environment and # /etc/security/pam_env.conf. auth required pam_env.so # [1] # In Debian 4.0 (etch), locale-related environment variables were moved to # /etc/default/locale, so read that as well. auth required pam_env.so envfile=/etc/default/locale # Standard Un*x authentication. @include common-auth # Disallow non-root logins when /etc/nologin exists. account required pam_nologin.so # Uncomment and edit /etc/security/access.conf if you need to set complex # access limits that are hard to express in sshd_config. # account required pam_access.so # Standard Un*x authorization. @include common-account # Standard Un*x session setup and teardown. @include common-session # Print the message of the day upon successful login. session optional pam_motd.so # [1] # Print the status of the user's mailbox upon successful login. session optional pam_mail.so standard noenv # [1] # Set up user limits from /etc/security/limits.conf. session required pam_limits.so # Set up SELinux capabilities (need modified pam) # session required pam_selinux.so multiple # Standard Un*x password updating. @include common-password

    Read the article

  • BIND split-view DNS config problem

    - by organicveggie
    We have two DNS servers: one external server controlled by our ISP and one internal server controlled by us. I'd like internal requests for foo.example.com to map to 192.168.100.5 and external requests continue to map to 1.2.3.4, so I'm trying to configure a view in bind. Unfortunately, bind fails when I attempt to reload the configuration. I'm sure I'm missing something simple, but I can't figure out what it is. options { directory "/var/cache/bind"; forwarders { 8.8.8.8; 8.8.4.4; }; auth-nxdomain no; # conform to RFC1035 listen-on-v6 { any; }; }; zone "." { type hint; file "/etc/bind/db.root"; }; zone "localhost" { type master; file "/etc/bind/db.local"; }; zone "127.in-addr.arpa" { type master; file "/etc/bind/db.127"; }; zone "0.in-addr.arpa" { type master; file "/etc/bind/db.0"; }; zone "255.in-addr.arpa" { type master; file "/etc/bind/db.255"; }; view "internal" { zone "example.com" { type master; notify no; file "/etc/bind/db.example.com"; }; }; zone "example.corp" { type master; file "/etc/bind/db.example.corp"; }; zone "100.168.192.in-addr.arpa" { type master; notify no; file "/etc/bind/db.192"; }; I have excluded the entries in the view for allow-recursion and recursion in an attempt to simplify the configuration. If I remove the view and just load the example.com zone directly, it works fine. Any advice on what I might be missing?

    Read the article

  • What the hell was THAT?!?

    - by Massimo
    My system is Windows XP SP3, updated with the latest patches. The PC is connected to a Cisco 877 ADSL router, which does NAT from the internal network to its single static public IP address. There are no forwarded ports, and the router's management console can only be accessed from the inside. I was doing two things: working on a remote office machine via VPN and browsing some web pages on the Cisco web site. The remote network is absolutely safe (it's a lab network, four virtual servers, no publicly accessible services and no users at all; also, none of what I'm going to describe ever happened there). The Cisco web site... well, I suppose is quite safe, too. Suddenly, something happened. Strange popups appears anywhere; programs claiming they're "antimalware", "antispyware" et so on begins autoinstalling; fake Windows Update and Security Center icons pop up in the system tray. svchost.exe began crashing repeatedly. Then, finally, after some minutes of this... BSOD. And, upon rebooting, BSOD again. Even in safe mode. Ok, that was obviously some virus/trojan/whatever. I had to install a new copy of Windows on another partition to clean things up. I found strange executables, services and DLLs almost anywhere. Amongst the other things, user32.dll and ndis.sys had been replaced. A fake software called "Antimalware Doctor" had been installed. There were services with completely random names or even GUIDs (!), and also ones called "IpSect" and "Darkness". There were executable files without an .exe extension. There were even two boot-class drivers, which I'm quite sure are the ones that finally caused the system to crash. A true massacre. Ok, now the questions: What the hell was that?!? It was something more than a simple virus! How did it manage to attack my computer, as I am behind a firewall and was not doing anything even only potentially harmful on the web at the time?

    Read the article

  • Hyper-V Machine drifts time all over, even with NTP

    - by MichaelGG
    Resolved The problem was Hyper-V on that machine. I removed Hyper-V, installed VMware Server, ran the same VM. Time sync issues went away (< 100ms difference after a day). My setup is like this: HYV1 - HyperV machine (non domain) - sync irrelevant AD1 - VM AD server on HYV1, sync'd to time.nist.gov. HyperV time sync off. S1 - Physical machine, sync'd to domain. S2 - Physical machine running HyperV, sync'd to domain. V1 - Linux VM machine on S2, sync'd to AD1. No HyperV integration. AD1 and S1 have fine sync -- stripchart shows less than 100ms difference. S2 drifts like crazy. Here's a bit of the stripchart against AD1: 18:33:22 d:+00.0010138s o:+05.4101899s 18:33:24 d:+00.0010138s o:+05.4319765s 18:33:26 d:+00.0000000s o:+05.4788429s 18:33:28 d:+00.0000000s o:+05.6089942s 18:33:30 d:+00.0010138s o:+05.7240269s 18:33:32 d:+00.0000000s o:+06.0421911s 18:33:34 d:+00.0081104s o:+06.5613708s 18:33:37 d:+00.0000000s o:+06.9096594s 18:33:39 d:+00.0000000s o:+06.8867838s 18:33:41 d:+00.0010127s o:+06.8936401s In 20 seconds, it drifted over a second. If I manually reset it to within 1s, within a few minutes it'll be back drifting about 2 seconds. Overnight it went from ~2s to ~5s. The Linux VM inside S2 has perfect sync with AD1. Here's the config: C:\Users\mgg>w32tm /dumpreg /subkey:Parameters Value Name Value Type Value Data ------------------------------------------------------------ ServiceDll REG_EXPAND_SZ %systemroot%\system32\w32time.dll ServiceMain REG_SZ SvchostEntry_W32Time ServiceDllUnloadOnStop REG_DWORD 1 Type REG_SZ NT5DS NtpServer REG_SZ ad01.mydomain ad02.mydomain C:\Users\mgg>w32tm /dumpreg /subkey:Config Value Name Value Type Value Data ----------------------------------------------------------- FrequencyCorrectRate REG_DWORD 4 PollAdjustFactor REG_DWORD 5 LargePhaseOffset REG_DWORD 50000000 SpikeWatchPeriod REG_DWORD 900 LocalClockDispersion REG_DWORD 9 HoldPeriod REG_DWORD 5 PhaseCorrectRate REG_DWORD 1 UpdateInterval REG_DWORD 30000 EventLogFlags REG_DWORD 2 AnnounceFlags REG_DWORD 5 TimeJumpAuditOffset REG_DWORD 28800 MinPollInterval REG_DWORD 2 MaxPollInterval REG_DWORD 8 MaxNegPhaseCorrection REG_DWORD -1 MaxPosPhaseCorrection REG_DWORD -1 MaxAllowedPhaseOffset REG_DWORD 300 I looked at the event log, and apart from warnings about sync (after it gets way out of sync), there's no other warnings. How can I go about troubleshooting this? It's the only machine that is having this problem. All the other machines (physical and virtual) are doing fine. Edit: To clarify: The VM (AD1) has integration turned off and syncs to time.nist.gov. AD1 is fine. It's the physical machine S1 that can't sync to AD1 and drifts all over. All the other physical servers are able to sync to AD1 just fine. Update So, it appears to be an issue of running the VM. The clock slips slowly with the VM off. Turned on, it immediately starts losing seconds. I swt the VM to only use half the resources, and that seems to have slightly mitigated it, for now. Thanks!

    Read the article

  • Windows Server 2008 R2 time keeps changing unexpectedly

    - by lmhost
    There is a serious problem with one of my servers time. At some random hours (5:59 localtime in the following example) it goes 1 hour back and keeps doing this again and again next times it reaches 5:59. Like this: 5:59-4:59 … 5:59- 4:59 and so on. It stays in the loop until I manually update its time. The server OS is Windows 2008 R2 Standard x64, acting as a standalone web server on the internet. time zone (UTC-05:00). system log: Information 4/13/2012 5:59:34 PM Kernel-General 1 None Information 4/13/2012 5:59:34 PM Kernel-General 1 None Information 4/13/2012 5:59:34 PM Kernel-General 1 None Information 4/13/2012 5:59:34 PM Kernel-General 1 None Information 4/13/2012 5:59:34 PM Kernel-General 1 None Information 4/13/2012 5:59:34 PM Kernel-General 1 None all of the above entries have same content: The system time has changed to ?2012?-?04?-?13T21:59:34.500000000Z from ?2012?-?04?-?13T22:59:34.500000000Z. Details XML View: - <Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event"> - <System> <Provider Name="Microsoft-Windows-Kernel-General" Guid="{A68CA8B7-004F-D7B6-A698-07E2DE0F1F5D}" /> <EventID>1</EventID> <Version>0</Version> <Level>4</Level> <Task>0</Task> <Opcode>0</Opcode> <Keywords>0x8000000000000010</Keywords> <TimeCreated SystemTime="2012-04-13T21:59:34.500000000Z" /> <EventRecordID>4060</EventRecordID> <Correlation /> <Execution ProcessID="4" ThreadID="80" /> <Channel>System</Channel> <Computer>********</Computer> <Security UserID="S-1-5-18" /> </System> - <EventData> <Data Name="NewTime">2012-04-13T21:59:34.500000000Z</Data> <Data Name="OldTime">2012-04-13T22:59:34.500000000Z</Data> </EventData> </Event> Any idea about what’s going on? Thanks

    Read the article

  • Providing DNS redirection to honeypot server for known bad domains

    - by syn-
    Currently running BIND on RHEL 5.4 and am looking for a more efficient manner of providing DNS redirection to a honeypot server for a large (30,000+) list of forbidden domains. Our current solution for this requirement is to include a file containing a zone master declaration for each blocked domain in named.conf. Subsequently, each of these zone declarations point to the same zone file, which resolves all hosts in that domain to our honeypot servers. ...basically this allows us to capture any "phone home" attempts by malware that may infiltrate the internal systems. The problem with this configuration is the large amount of time taken to load all 30,000+ domains as well as management of the domain list configuration file itself... if any errors creep into this file, the BIND server will fail to start, thereby making automation of the process a little frightening. So I'm looking for something more efficient and potentially less error prone. named.conf entry: include "blackholes.conf"; blackholes.conf entry example: zone "bad-domain.com" IN { type master; file "/var/named/blackhole.zone"; allow-query { any; }; notify no; }; blackhole.zone entries: $INCLUDE std.soa @ NS ns1.ourdomain.com. @ NS ns2.ourdomain.com. @ NS ns3.ourdomain.com.                        IN            A                192.168.0.99 *                      IN            A                192.168.0.99

    Read the article

  • AWS EC2 instance not pingable or available in browser

    - by Slimmons
    I've seen this questions asked other places, but now I've run through every fix proposed in other questions so I'm re asking it here, in hopes that someone will have a different solution. Problem: I have a EC2 instance, and I can ssh into it and work on it, and I have a Elastic ip set to it. I am unable to ping this machine, or log in to it using my browser. Solutions mentioned and tried: service httpd start i. response I get is "unrecognized service" ii. when I run apache2ctl -k start, it shows "httpd already running", so I'm assuming httpd is not the problem, it's just possibly named something else because of apache2, or for whatever reason. I went into EC2-Security Group- Default (which is the one I used.)-inbound, and everything there is set up correctly (I'm assuming). There it shows 80(HTTP) 0.0.0.0/0. 443(HTTPS) 0.0.0.0/0, and various other servies with their ports and 0.0.0.0/0 next to them. I also enabled a rule for enabling ICMP Request All on 0.0.0.0/0 temporarily for testing purposes I've tried disabling the iptables with "service ufw stop" Just in case I'm doing something really stupid, because I'm not all that used to connecting to web servers that I've spun up, I'm typing in the address to the machine into the url like this (assuming my ip address was ip.address). i. http:/(slash)ip.address/ ii. ip.address iii. https:/(slash)ip.address/ iv. ip.address/webFolderName/ v. http:/(slash)ip.address/webFolderName/ None of the attempts worked, and the only thing I haven't tried that i've seen is to start wireshark on the machine, and see if the requests are reaching it, and it's just ignoring them. I'm not sure I want to do that yet, since A). I'm not 100% positive how to use wireshark without the gui, since it's the only way I've ever used it (I really should get used to it in terminal, but I didn't even know you could). B). It really seems like I'm missing something simple in getting this to work. Thanks in advance for any help.

    Read the article

< Previous Page | 356 357 358 359 360 361 362 363 364 365 366 367  | Next Page >