SLOW WAN replication

Similar Questions

• Turn off the site without cloud or WAN backup options

  Just started with a job with the new company, which is still tape backup old school that need replaced.   I used Appassure in previous environment with WAN replication which was awesome.    I want to offer a solution to the new company but a mandate corporate requirements;

  1. daily backups offsite

  2. No WAN connectivity off site.

  Now they have daily mail which incorporates the previous nights Ribbon and stores them in a vault of company off-site 3rd part.

  My other projects this year are to virtualize them and also offer a DR plan for all their systems.  It's a pretty small environment with 14 physical servers and less than 10 TB of data.

  Although second requirement makes life quite difficult to achieve (1).

  Guess you'll have to try to figure out why this second requirement and change it (i.e. No WAN except at the site of the replica) or continue to use a type of cone / disk backup solution.

  You could possibly use Appassure and create regular archives to 10 TB NAS that you take daily off-site :).

  I think your options are limited, I'm sorry, but the band will probably be your only real option.

• UTL_SMTP becomes suddenly slower

  Hello world
  
  We use the package UTL_SMTP to send mails to our applications. There is no if long, the e-mail were really fast. Now, it takes more than 10 seconds to send 3 emails with no attachment, it is not acceptable. You know a way to send mail faster?
  
  I read somewhere that in Oracle 10 g, we should use instead of UTL_SMTP UTL_MAIL? Would it be spead my generation of mail?
  
  Thank you!
  
  Bodart!

  leinadjan wrote:

  I think I found the culprit, it's the connection to the SMTP server. Each team of the procedure's appeal, we will open a new connection and shut us down. And the average connection time is 5 seconds. This three times and we have to wait 15 seconds.

  Does not seem reasonable. A tcp socket connection must be faster - if it's in the (slower) WAN infrastructure.

  The SMTP server could do DNS inverted for logging purposes. It accepts a client connection. The IP address of the client is available - but instead to use this only for purposes of logging, it can search the host name of this IP address. And then to use the host name for logging. (she also can use same basic access control based on the host and domain names),

  This is called a search reverse DNS and can translate by the initial client connection much slower.

  But I think I have must at least try to share connections for all mail and it will be a small gain of speed.

  You can't share a socket connection in sessions of PL/SQL. So, if the 123 session creates an SMTP connection, only he can use this connection.

  What you can do is put up a server to send mail from PL/SQL database. A PL/SQL package can use every method available in Oracle for the other sessions to communicate with IPC (Inter Process Communication). E.g. database of pipes, or queues or even just a simple table SQL vanilla.

  This package can create an SMTP connection, keep alive and remove all messages it receives other Oracle sessions. A corresponding message is transformed into a message and the SMTP connection is used to send it.

  This package can be run as a background using DBMS_JOB/DBMS_SCHEDULER process.

  Some methods of IPC supports multiple processes on the server - side cases this approach can be resized in something like 5 copies of this mail server code running. This supports then 5 simultaneous requests for sending mail. Kind of like the concept of shared/MTS server in Oracle.

  UTL_SMTP.close_data (v_connectionsmtp); What is the command that end the creation of my mail?

  Fix. This sends a period and a CRLF to the server to indicate that the e-mail full has been send.

  The server responds with a digital code (which UTL_SMTP must check the confirmation), then the server expects the next SMTP command from your code.

  I want to manage the connection manually, open it when it is null and close at the end of my script...

  Don't know what mean you "+ open when it's null +"?

  If you do not create a mail sending server in the PL/SQL code, it should look something like the following:

  connect to SMTP./mail server
  send HELO command
  while PLSQLServerIsActive loop
    if MessageOnQueue then
      DequeueMessage
      send MAIL FROM command
      send RCPT FROM command
      send DATA command
      send e-mail
      send period + CRLF (end-of-mail)
    end if
  
    if TimeIdle > 60 seconds then
      // tell the SMTP server that we're still here in order
      // to prevent it from timing us out and closing the
      // tcp connection
      send RSET command
    end if
  
    // shutdown when commanded to
    if ReceiveServerShutdown  then PLSQLServerIsActive := false
  end loop
  send QUIT command
  

• (With WAN replcation) failover cluster

  Hello
  We have 3 different groups with caches which are synchronized via the WAN replication. Now, we are planning to implement the cluster failover in which if a cluster is unavailable or unstable client (uses TCP extend) would failover to another cluster.
  
  Now, it is easy to configure the client by mentioning same cache with different names in different groups. But what makes configuring complex and error prone, especially when the number of clusters, but also caches should increase.
  
  So what I'm looking for is a way I can change the schema of underlying caching for a given cache name. So that I can point dynamically to the other cluster without having to restart my client.
  
  Or if there a better way to do it, clues are welcome :)
  
  
  Thank you
  Kishore s.
  
  Published by: user10737736 on July 14, 2009 05:31

  Hello
  I don't know if that's what you're looking for, but I have in past used ConfigurableCacheFactory to solve a similar problem.
  Essentially, I create a client to a cluster using ConfigurableCacheFactory connection, I created by instantiating a DefaultConfigurableCacheFactory by the way of an XmlDocument which essentially my client configuration file for consistency. Of the ConfigurableCacheFactory, you can create a CacheService and connect according to normal.
  Now if you have a power failure you can create another foregoing a cluster, closing the existing service of course.
  Kind regards
  Nick

• BEFSR41 and reduced connection speed

  I got cable high speed last week, and with it connected directly to my PC, I got amazing 28-33 Mbps connection speeds.

  So I plugged the router today to coonect mine and my wife (via Cat5) PC to cable and now I get only about 5 Mbit/s.

  The router choking off the coast of flow? Do I need a different model that can handle the speed?

  What the BEFSR41 version do you have?

  This problem of slowness "WAN to LAN speed" has been reported with version 1 and version 2.  Version 2 seems to be limited to about 7 Mbps.   Version 1 is slower.

• Cannot access shared files whne connected to the server

  Original title: dfs namespace access

  We have customers from windows 7 on windows 2008r2 domain. We have a couple of file hosting a dfs servers space names with file shares.

  This server hosts a common \\namespace\fileshare that is still available however 2 other \\namespace\share1 & 2 return cannot access this share when you are not connected to the server. This does not affect any clients, but affects enough that it becomes big problem being the only customer resolution of reconstruction. I tried adding and removing updates. I am at a loss what to try next.

  Any suggestion would be great.

  Thanks Mick

  I found the solution myself after a lot of googling and journal of trace events.

  After you configure wsus, it appeared that our Windows 7 clients suddenly received file sync (small green icon in the taskbar) couple this with an intermittent slow wan link. The fix was all right before you disable offline files. We never wanted their first place very frustrating, but we're still learning.

  It doesn't seem to affect all clients which is what got us thrown away but hey its works now.

  Thanks Mick

• VDR or Veeam

  Hello

  For the two sites on the Wan replication, what are the parameters to compare between two solutions, including the price

  Thank you

  The first and most important differentiating factor is that Veeam replication tasks allow you to have multiple points of restoration/restorations. It is a function of essenital, cause like ' good/valid', any data the viruses/corruption/loss of WHAT VM source is immediately replicated to target VM. And I think you would agree that it would be very difficult to detect such problems fast enough in the order of failover to replica VM before the next replication cycle comes into play.

  In addition, because you want to replicate on the WAN link, I guess you must also features traffic compression and limiting traffic and replication tasks Veeam provide this functionality.

  I hope this helps!

• How can I assign a subnet on VMPlayer 4.0?

  I followed this tutorial on how to extract the network editor which is not installed by default:

  http://www.technazgul.com/2011/01/advanced-network-configuration-using.html

  The question, I am running is the VMPlayer 4 version has less features compared to the old version, and I do not know how to assign a subnet to a virtual computer.  Here is a screenshot of what the interface looks like to my test network:
  
  I'm trying to simulate a network with two domain controllers in different sites that are connected by a slow WAN link.

  If I understand correctly, you want to assign an individual VMnetX to your VM. With VMware Player, you can do that by editing the virtual machine .vmx file manually (see for example Re: several network adapters and vmnetcfg works not properly)

  André

• To defragment or not?

  We have a particular server that hosts several Synergex databases.  There are also a number of flat type files that get deleted/copied/added every day.  According to Windows Server 2008 R2, the disk is highly fragmented.  This server exists on the network SAN Equallogic PS6000XV.  It is part of the RAID 50, where it has its own dedicated volume.  The VMDK is thick-set service.

  When he was a physical server and the data was hosted on Server 2003 with a local RAID-5, we had installed Diskeeper and could run a defrag every week.  It would be a good idea now?  Looking through the forums, I've seen arguments in two ways.  Most are quite dated, so I wondered what was the current consensus.  I see three ways to accomplish this:

  1 defragment the SAN, which is even possible?  Does anyone know if Equallogic has a tool for that?

  2 defragment the VMDK?  Again, it is like Voodoo?

  3 defragmentation of the file system in the VMDK - using tools like Diskeeper disk or something?

  or

  4. just leave well enough alone, because Windows doesn't know what he's talking about.

  When he was physical, we saw significant performance increases after the first defragmentation.  In both cases, it works a lot better in a virtual environment with fewer resources than what he has done as a physical server.  Virtual stuff never ceases to amaze me.

  We have a particular server that hosts several Synergex databases... This server exists on an Equallogic PS6000XV... part of RAID 50, where it has its own dedicated volume.  The VMDK is thick-set service.

  The expected outcome of defragmentation, it's what to consider - as André, there is no certainty about what made the EqualLogic with blocks and in any case, even if there was, because the volume is spread over 14 pin, what exactly is sequential access even when?  The EqualLogic provides storage for something else?  Databases the most tend to show a random IO load because they meet (and scattered) data tables so how it is organized physically just not important really.
  Positioning system EqualLogic data is skilled enough - for example, if in the future an SSD array had to be added, the system could migrate hot blocks in this table.
  Replication is also extremely important.  IIRC, EqualLogic use the block of tracking changes with a block size of 64 MB.  So at the end of the replication interval, all 64 MB blocks that have been affected by a write will be transmitted over the Wan - clearly defrag will touch a massive number of blocks and potentially make this impractical.
  There are also a number of flat type files that get deleted/copied/added every day.
  In the same sense, it may be necessary to consider.  For example, you might consider such temporary space running in folders with NTFS compression enabled. If these are text files that can be very effective to reduce the disk space required (and therefore the number of blocks received for WAN replication).   Another option would be to run these from the temporary space provided by a server of files on a volume that is not replicated.
  Hope that helps!

• ESXi: SAS Raid 1 Boot w / SATA for backup?

  Situaltion:

  • There is currently no effective backup strategy.  Download that data over a slow WAN link can be used to restore or install timely.

  • 2 units of each SCSI 146 GB for $100 / month.  2 each 1 TB SATA disks for $100 / month, provider supplied NAS 300 GB for $150 / mo.  Seller provided iSCSI for 300 GB to $200 / month.

  • Cost per GB SCSI is not yet close to that of SATA.  Two 146 GB drives would be very marginal.

  • I intend to load balancing and cluster two sites on servers with something like CARP.  I am open to any suggestion using NAS and iSCSI.  The main things that matter are mysql used by http, http and email.  Sites are dynamic and use PHP which requires the session data.

  Mission:

  • Store copies of virtual machines, store ISO images, make backups of the operating system and perform backups of application server to provide access to the required software and backup and highspeed images.

  • Retain the ability to balance the load and 2 cluster sites.

  Equipment available:

  • Two ESXi servers run SuperMicro X7DWU motherboard (Compatible with ESXi) Adaptec RAID 31605 controller RAID-1 running. (http://www.vmware.com/resources/compatibility/detail.php?device_cat=io & device_id = 9051)

  • Each server has two network cards integrated, private one on the network, eth0 and the other on the public network eth1.  (Intel Corporation Intel 82575EB Gigabit)

  • Other controllers on the motherboard are: controller RAID Intel ESB2 SATA 3.0Gbps.  Others got this controller to work with ESXi on the same motherboard on this forum.  (http://communities.vmware.com/message/891637#891637)

  Execution:

  I need your ideas to develop a good strategy.

  1. Yes.  You can assign to ESX directly and create a VMFS out of it, and then give that space for the virtual machine that you use for the backup.  You have to be careful and make sure that go you through the HCL (vmware.com/go/hcl) to get a SATA controller supported.  Not all controllers will work and be recognized by ESX.

  2 Yeap.  You need not necesarilly CYGWIN for NFS, there are tools of $ M that can be loaded as well Windows services for unix, either will endeavour to provide an NFS server.  The downside is you need a windows license, but the customer is handy to have.  You can use a virtual linux machine, and even if you can't use the vi client, you can use the remote cli tools that are available for Linux.  Vima (VI Management Assistant) unit is a device based on linux with the remote cli installed, might want to take a look if you have not already.  You can use samba on it and be able to map a drive as your windows machines.

  3. Yes, absolutely.  If you have an NFS server, you can also use it as a data store NFS for the 2nd Server.

  4 VCB is very convenient and it will allow you to skip creating a server and the export of the step of the vm, but you will still need to configure vcb proxy, which will need a windows server.

  -KjB

• Exrteme AC WAN slow airport

  Hello

  I encountered the following problem:

  If I test my internet connection directly with PPOE connection since my MacbookPro get 900MB or more download speed.

  If I'm testing with the router (with the same cable) I get only 250 MB speed.

  Any ideas what might be the problem? As far as I know this model has 1gbit WAN port.

  Thank you very much

  Zoltan

  Any ideas what might be the problem? As far as I know this model has 1gbit WAN port.

  Yes, it is... but NAT in router is not powerful enough to manage gigabit connections...

  I'd say you're still shy of the number... I expect about 400-450Mbps as the best, that the processor can handle without hardware NAT.

  The problem is firmware Apple removes material NAT... and without it the speed decreases considerably... but you probably other things running that is causing slow downs.

  For very fast router flow watch list in the table Smallnetbuilder.

  http://www.SmallNetBuilder.com/tools/charts/router/view

  Apple Airport extreme is in the list... a long long way down. Place about 80... with prices upwards... the design is now 4 years old and simple fact is that they have been exceeded on all fronts.

• EqualLogic remote replication an auto-montage cascade, can link a replica for the slower berries?

  Hello, I was wondering if anyone had an answer to that:

  We have recently installed 8 bays EqualLogic (4 x 6100 x + 4 x 6100F).

  We have two servers and have set up a SAN array in each part of the machine consisting of a simple storage pool (2 x 6100 (RAID50) + 2 x 6100F (RAID6)) in order to use the auto-montage cascading.

  I played with the replication distance, and I can't seem to find a way to link a replica to a RAID type or a member. Ideally I would like to bind all replicas to our tables more nits that are running in RAID 6.

  It doesn't seem that this is possible, this is possible by having one or more slower berries in a different pool and replicate to this pool instead. The reason why I was initially resistant to this idea was that I wanted to be able to exploit the performance of all the bays via auto-etagement, as opposed to spending tables for replication that would do anything for most of the day.

  Thank you

  Hello

  OK, you can decide what to put in common the delegate space is, but not which Member or RAID type.

  Kind regards

• Slow flow on MPLS VPN WAN

  Anyone have any ideas why a portion of the traffic is slow as it passes through a VPN MPLS WAN. My FTP copies are fast but copy all windows or windows file transfers are slow. Copies of windows are about three times slower as the FTP transfers. Can be optimized on routers or switches?

  Hello

  Thus, all transfers are done with CIFS are slow and other then CIFS are ok?

  All transfers are between XP/7 and servers (before 2008)?

  Please take a look at http://bit.ly/rkh9IM

  CIFS (or SMB) prior to the 2008 version is slow by definition as it can not cope with very good latency. Other protocols such as HTTP and FTP run much smoother.

  When you run Server 2008 (or better) combination with Windows Vista (or better) should solve some of your problems as it can using SMBv2.

  What actual speed is your order on the MPLS and what is the maximum transfer reached between server and workstation?

  Best regards, G.

• RV180 WAN/Internet very slow speed via WAP4410n (Port RV180 speed problem)?

  Hello

  I have a base set up on the RV180, no etc. VPN running 1.0.5.4 Firmware (latest) and use a browser of WAP WAP4410n plugged into a LAN of the RV180 port. I have a SG100D Gigabit Switch also connected to a port RV180. My WiMax ISP provides 12 MB.

  When I leave speed Port on the RV180 & WAP4410n on Auto neg. They hook at 1 GB and I have ONLY 2 MB Internet speed on wireless clients with no reported ignored packets on each device, when I change the port to 100 MB on the WAP4410n Rx. speed speed is normal at 12 MB!  The RV180 also indicates the speed of the LAN port for 1 GB... it is obviously wrong!  Obviously, forced to run the LAN of WAP port to 100 MB so I can get the ISP... normal speeds is limitation all my wireless to wired LAN traffic (and vice versa).

  I tried several configs RV180 ie. Simplified of bare bones, exchanged cat 5th etc. etc. WITHOUT change.  With another router, there is no problem with the port of WAP4410N at 1 GB and ISP speeds are normal. ISP speeds are also normal via the Gigabit switch to other boxes, BUT if I plug the WAP4410N in switch WAN speed is 2 MB if the WAP4410N is set to 1 GB port speed!

  With the RV180 unable to define a manual port 100 MB speed (see 1.0.5.4 Firmware release notes), that strongly leads me to believe that it is a problem of data Port RV180 specific WAP WAP4410N somehow.

  Their are many positions reports slow Internet with the RV180, but none of them have a solution.  Ideas, any help would be greatly appreciated that Ive lost too much time already trying to solve this router ÷ # $$ / / / &. As this is my first post I hope I posted in the appropriate forum.

  Thank you... Rob.

  Thank you for the update.

• the WAN connection becomes too slow after you have configured the VPN (Site Site)

  Hello

  I have two branches connected via WAN (MPLS) connection using two 2921 routers.the connection is 2 M.

  I set up a VPN between these two sites, but after the connection has become very slow.

  y at - it something I can do to speed up the speed of connection.

  VPN proposals are:

  Proposals of the phase 1: 3DES, pre-shared,.

  Phase 2 proposals: esp-3des esp-sha-hmac

  I don't think that lower levels of security proposals will add a lot to the speed...

  Hi Marc,

  one thing you should definitely is a hardware encryption go if you do not already tht, it also reduces the load on your cpu

  You can try other things is play with mtu, according to your line mtu and what applications are mainly used. try setting the mtu of at least 60 odd bytes lower than the mtu and also sometimes server line recommended mtu settings like server many have obligation to mtu to 1300 or 1400, if that's not it can cause a lot of re transmissions, you can also try fragmentation before encryption

  http://www.Cisco.com/en/us/docs/interfaces_modules/services_modules/VSPA/configuration/guide/ivmvpnb.PDF