Introducing Nomad 6.3
Working with Nomad
Technical support for Nomad
Relicensing Nomad client
Resolving common issues
Resolving content integrity (hash checks) issues
Resolving dynamic pre-caching issues
Resolving hard link issues
Resolving Nomad client health issues
Resolving Nomad Dashboard issues
Resolving Nomad peer election issues
Resolving PBA issues
Resolving peer copy over HTTP or HTTPS issues
Resolving remote differential compression issues
Resolving slow content transfer issues
- Known issues
Installing and working with Nomad
|Nomad is not able to download from CCMTOKENAUTH_SMS_DP_SMSPKG$ content location virtual directory on Distribution Point, which is created when Distribution Point is configured to communicate over HTTPS starting CM CB 1810.|
Enhanced HTTP was first introduced in CM1806 as a pre-release feature and it became a standard feature in CM1810. It allows administrators to use secure communication between clients and DP without the need for PKI-generated certificates or a Network Access Account.
Nomad currently can’t use the new Enhanced HTTP features when downloading from the DP, but it will continue to work when there are still Network Access Accounts available to allow it to authenticate non-Domain joined machines (such as those in WinPE and Workgroup machines).
While in Windows PE though, whether using Enhanced HTTP or HTTPS, Nomad can fail to download content from the DP. A solution for this issue is to add a Run Command Line step in the Task Sequence that runs the following command to define the DpNotAvailableCodes registry value:
This needs to be after Install and Configure Nomad in Windows PE step.
There is an update to the Install and Configure Nomad in Windows PE step in Nomad 7.0 and in an upcoming Hotfix for Nomad 6.3, which includes the configuration of the DpNotAvailableCodes so that the solution above is no longer necessary. You would still need the aforementioned Network Access Accounts.
We are working with Microsoft to determine the best approach to enabling download from the Enhanced HTTP virtual directory. You can upvote our request at https://aka.ms/AA63jc3
|Nomad download halts for some time if block size is different on a machine acting as a master.||Nomad may not be able to download whole content from the master and stuck in a loop for some time if the master has incomplete content and different block size.|
|Nomad won't support Office 365 jobs to download from Internet||Nomad won't be able to download Office 365 update if their data source is some internet URL.|
|Different block sizes on DP and Clients may leads to RDC failure.||RDC may not work properly if distribution point and client has different block size. To fix this, have same block on all distribution points and client.|
|Client health report would get exported in .xlsx format even when PDF/Word format is selected.||Client health extended view reports would not get exploreted in desired file format. When user choose pdf or doc format, the report always gets exported in xlsx format.|
|Nomad would fail to download content when client switches from Intranet to Internet DP having higher content version.||Nomad may fail to download the content when switching from Intranet to Internet DP and Internet DP has higher version of content than required.|
|Nomad fails to use already precached content in secure format even if same content is requested in original format.||Nomad won't convert the secure format content into the original format on the client side and downloads the content again.|
|Nomad settings on SCCM client policies don't replicate between CAS and primaries||Nomad settings don't replicate between CAS and Primary site, the user has to manually do settings on both the sites.|
|Nomad won't handle SUSPEND by CM when in PauseJobQueue state||When Nomad JobQueue is paused using Content Pause and during the pause if CM sends suspend for a job then Nomad won't honor that instruction and would continue to run that job once JobQueue resume.|
|Nomad sometimes fails to disqualify the master machine.||During P2P download if master becomes unavailable, peer machine sometimes fails to send disqualify message to master and remains stuck for long retrying to connect to same master machine.|
|Nomad won't do recycle DP in WinPE.||Nomad in WINPE has limited set of registries and it does not have Recycle DP related registries. To fix this, Please create DpNotAvaiableCodes registry in WINPE.|
|Nomad logs shows download percentage more than 100% and register 99.99% to AE.||Nomad may show incorrect download percentage if LSZ file gets corrupt on the client side.|
|NomadBranch GUI does not work when File sharing is connected with Remote machine||When NomadBranch GUI is installed on machine and connected to remote machine then it may not show any download progress and remains in paused state,|
|NomadAdminUIExtension always give "stopped working" error but installs successfully on Windows Server 2012.||Installing NomadAdminUIExtension on Windows Server 2012 would give error "stopped working" however installs successfully. Users just have to say OK to error prompt.|
|Incorrect protocol status download reported in the ||The |
|Nomad crashes on service start-up||This happens while during the license verification process where Nomad looks for an instance of an obsolete WMI class.|
|The correct installer may not be picked up by the Install Nomad custom task sequence action if there are multiple installers in the content source for the Nomad package.||The Install Nomad custom task sequence action randomly selects a Nomad MSI if there are multiple installers in the content source for the Nomad package. We recommend having only one installer in the Nomad package source.|
|Existing Installation folder not detected by NomadBranch-x64.msi||Run the command-line with the installer property or apply a transform.|
|Latest deployment download progress chart shows incorrect results for user collections.||Reporting is incorrect in the Nomad dashboard for user collections when the same user downloads content from two or more machines. This event is a rare occurrence.|
|Job fails if disk space is low when compress/encrypted content is uncompressed.||For compressed/encrypted content, Nomad reserves disk space equivalent to size of compressed/encrypted content only before download. When disk space is low the, job will fail when it uncompresses the content in the Configuration Manager cache.|
|Status messages are not sent when Nomad runs in WinPE mode||When Nomad is installed in WinPE and starts downloading packages, send status messages are not sent to CM server as it does not have SCCM client.|
|Peer-to-peer transfer fails if one of the machine name is longer than 50 characters||When on a subnet a machine1 has hostname longer than 50 characters with content A and machine2 requests content A, P2P transfer fails and content will be downloaded from DP|
|The Install Nomad task sequence action randomly picks an MSI for installation||During the task sequence execution the Install Nomad task sequence action randomly picks an MSI for installation from package staged which containing multiple MSIs.|
|The Apply button is not enabled when parameters are specified for the Install Nomad task sequence.||Additional settings for the Install Nomad task sequence action fails to apply if specified directly|
|Attempting to downgrade nomad using SCCM puts the machine in such a state that any further upgrades would fail.||When a downgrade is attempted using a ConfigMgr job, the rollback doesn’t happen cleanly and a few traces are left in the system leading to failure in future upgrades. To fix this, please contact 1E support.|
|Manger 2012 distribution point is installed on a Configuration Manager 2012 primary site which stops Nomad Branch using RDC.||The following workaround resolves this issue:|
|Log files may contain a large number of error messages if a mixture of encryption methods is used for Nomad communications.||It is possible to configure different installations of Nomad to use different encryption methods. Nomad 6.0 supports two options: the default encryption method and a new FIPs compliant encryption method. The error messages in the log files indicate that Nomad communications are being impacted and the fault should be rectified.You should ensure that all the installations of Nomad are set to the same encryption method. |
This state is liable to occur if hybrid installations are allowed as no-encryption is no longer supported in Nomad 6.0.
|Clients may not be able to download content using HTTP/HTTPS from the Distribution Point if the content URL contains hidden segments or double-escape sequences||This issue is specific to Distribution Points running IIS 7.0 and above and may occur when the requested content URLs contain hidden segments or double-escape sequences. When this happens the client displays HTTP error 404.8 (Hidden Namespace) and 404.11 (Double Escaping) in the NomadBranch.log file while processing the content.|
To work around this issue, the IIS Request Filtering settings on the DP should be modified to allow download of such content. Please refer this link for details on how this can be configured.
|Using Nomad on multihomed machines results in extraneous network traffic||Please contact 1E Support for a detailed workaround if you have multihomed machines in your environment.|
|Disk Full error retries with Configuration Manager Client||If the Nomad job returns with the error Disk Full, the Configuration Manager Advanced Client regards this as a non-fatal error and will retry the deployment on a scheduled basis. This is by design.|
|Nomad Provider Settings overwrite Third party Provider settings in Task Sequences||When making changes to Nomad Properties for a Task Sequence the settings are written to all the Task Sequence's reference Packages. This does not take into account any third party configurations that may already be in place. To avoid this you should set Nomad Provider settings for each of the reference Packages individually.|
|Nomad encounters problems when WebDAV is used to connect to a server||Use of WebDAV is not supported by Nomad.|
|Uninstalling and reinstalling Nomad results in unverified cached data||When uninstalling Nomad, cache content is not deleted as files may still be required (for example for Windows Installer repairs). If you subsequently reinstall Nomad on the same machine, Nomad will enumerate the existing cache and will be aware that it has cached content for certain packages. It is not possible to verify that the package content has not been altered in the interim period, therefore these are referred to as zombie packages i.e. they have returned from the dead.|
To manually force verification of the cache after upgrading Nomad or uninstalling and reinstalling Nomad, please use NomadBranch.exe -activateall command
|Nomad crashes while downloading an application in original content having source path greater than 290 length|
When the source path of an application is too long e.g. more than 290 chars and Nomad starts downloading application in original format i.e. format 0 then Nomad crashes in between but successfully downloads the content and passes the hash check as well.
Workaround: When a package has source path greater than 255 characters in length, use nomad secure in order to avoid any crash or mishandling.
Disabling TLS 1.0 on ActiveEfficiency Server or clients breaks pre-caching and other features.
Disabling TLS 1.0 on ActiveEfficiency Server would cause the following issues:
Resolution: Please install the latest hotfixes for NomadBranch client and Nomad AdminUI Extensions.
|Dynamic packages in Task Sequences sometimes remain in Installing state.||When SMSTSNomad is used to download Dynamic Packages, through using the Set Nomad as Download Program action in a Task Sequence, they may remain in Installing state.|
|Download using Nomad during a Task Sequence occasionally reverts to using the default Configuration Manager download process.||When SMSTSNomad is used in a Task Sequence by using the Set Nomad as Download Program action, certain error conditions that would normally be translated to success codes by the list held in the SuccessCodes registry value are instead being returned as failures, thereby causing Configuration Manager to revert to the use of BITS for the transfer when those errors occur.|
|Task Sequence downloads may failover to using the default Configuration Manager download process if any errors occur during the Nomad download.||Due to a problem in Configuration Manager the Always return success option in the Set Nomad as Download Program action is not working. This means that downloads may failover to using the default Configuration Manager download process. You will need to regularly check the Nomad logs to check that this is not occurring in order to address any issues that may have arisen.|
|Content may not be downloaded after an OS has been applied to a machine.||If SMSTSNomad is used after an OS has been applied to a machine it will not use multiple DPs. So if download from the first DP fails it will not failover to the next DP in the Configuration Manager list.|
|The Nomad properties dialogue in the Configuration Manager console does not save changed values when focus moves to another tab||Where the Nomad Admin UI extensions are installed, in any of the dialogues in the Configuration Manager console where the Nomad properties appear as a tabbed pane (e.g. in the properties of a package), if a Nomad property is modified but the "Apply" button is not clicked before moving to another tab, the value is not saved and reverts to the original value.|
The workaround is to save the values by clicking the "Apply" button before moving to another tab.
|Certain issues with Applications in Configuration Manager 2012||Due to the Application model used in Configuration Manager 2012 the following features do not work when using Nomad, Applications and Configuration Manager 2012:|
|A failing task sequence step can be repeated when run by a Configuration Manager 2012 SP1 client on Windows 7 or Windows 8||If a task sequence step fails (and "Continue on error" is not set for the step) and aborts a task sequence, the task sequence engine may incorrectly restart itself after a few seconds and repeat the last, failed, step. This has been seen to happen when "Peer Backup Assistant: Nomad PBA Data Store High Availability" fails, but it can also happen with non-Nomad steps. Repeating a TS step that fails is usually harmless.|
|An exception may be seen when Nomad Admin UI extensions are upgraded and Task sequence editor is immediately launched.||An exception may occur when Nomad Admin UI extensions are upgraded and Task sequence editor is immediately launched. This problem does not usually occur and can be avoided by launching the Task sequence editor after 30 minutes once the upgrade has completed.|
|Nomad won't download Express updates.||Nomad won't be able to download Express Updates, it will handover the express update job back to CM.|
Windows 7 and Windows Server 2008 R2
|Nomad peer running Windows 7 or Windows Server 2008 R2 transferring a file via SMB that is larger than system memory degrades in performance||There are a number of known problems with the Windows 7 and Windows Server 2008 R2 implementation of SMB, particularly when many I/O operations are performed on a share. See for example MS KB article 2582112. Older versions of Windows (e.g. XP) and later (e.g. Windows 8) are okay. A primary feature of Nomad of course is that it transfers files in discrete blocks in order to throttle its network bandwidth usage, even in P2P transfers.|
Large packages whose constituent files are all smaller than the system memory are not a problem.
There are a number of possible workarounds. For example:
|Nomad Multicast is disabled by default on Windows Vista 64-bit and Windows Server 2008 64-bit if a MADCAP scope name is used||The MADCAP client API implementation on Windows Vista 64-bit and Windows Server 2008 64-bit to enumerate scope names is faulty which can cause the Nomad 64-bit agent to quit unexpectedly on these operating systems if Nomad multicast is enabled. Hence, the multicast settings are automatically ignored if MultiCastMADCAPScope specifies a scope name e.g. SMSNomad, and Nomad multicast support gets disabled. However, using IP addresses for MultiCastMADCAPScope (i.e. a value of the form <MADCAP_ServerIP>,<MADCAP_Scope_IP>,<TTL>) is safe and will not be disabled.|
This issue doesn't affect Nomad 64-bit agent on any other supported operating systems (e.g. Windows 7 or Windows Vista 32-bit and Windows Server 2008 32-bit) and Nomad 32-bit agent.
|Multicast encountering problems on Windows 7, Windows 8, Windows Server 2008 R2 and Windows Server 2012 computers||The advanced Windows firewall on these computers by default blocks any inbound traffic from the MADCAP server. This issue is related to the MADCAP scope used for Nomad multicast.|
|Nomad multicast has problems multicasting on certain VPNs||Nomad may not be able to use multicast via a private enterprise server using a VPN connection on a client computer that is running Windows 7, Windows Server 2008, Windows Vista, or Windows Server 2008 R2. There is a hotfix available from Microsoft that attempts to address this issue: http://support.microsoft.com/?kbid=981637|
|Nomad may not check available disk space in a timely fashion during a Multicast transmission||If a PC joins the multicast session mid-way through, it immediately starts to cache data locally and is not aware of the Package size. This is by design. In environments where disk space is critical this issue is mitigated by:|
|“Slow sockets” log message||This error message occurs during distributions using multicast if some datagram packets are lost. Some packet loss is to be expected, especially if a network traffic peak occurs, Nomad is designed to fill in the lost data at the end of the distribution. If excessive logging of this error message occurs and the statistics are returning a low percentage of data multicast, your network infrastructure may not be suited to efficient multicast. This can occur if you have an infrastructure with mixed network speeds for example.|
|Using MADCAP over very slow links with pre-cached data||In scenarios with very slow WAN links (i.e. <256kbps) and pre-cached package data, peer copying is used instead of local multicast. If this occurs the following warning message will be displayed in the log file.|
WARNING: MADCAP_Request for PACKAGE_ID discarded $$<MCastControl><datetime>
This happens due to the time taken to get a MADCAP address over the WAN. Package data is still shared at the branch and only copied once over the WAN successfully, however local network traffic will be higher as peer copying is used instead of local multicast. MADCAP performance can be improved by specifying the MADCAP server IP address in the Nomad MULTICASTMADCAPSCOPE registry value.
|Multicast continues after deployment has expired||If the deployment expires during a multicast transmission, the transmission will continue.|
|Multicast is not supported upon FAT file systems||Nomad with Multicast is not currently supported on FAT file systems. This may be addressed in a later version, pending customer feedback.|
SMSNomad.exe as a program command-line
|Using “Start in” Configuration Manager Program property||SMSNomad.exe assumes the working folder is the Distribution Point. If a “start in” location is specified as an Configuration Manager Program property then SMSNomad.exe is unable to locate the package source and fails to distribute software.|
|Nomad registers with ActiveEfficiency platform using network interface other than primary interface on multi-homed machines||On multi-homed machines Nomad may register with a network interface other than the primary network interface. This may result in the machine becoming unavailable for single site download to other machines in the same site since it has registered itself with a network interface other than the intended one.|
|Zero byte content does not register any intermediate content value with ActiveEfficiency on incomplete download.||If content having size of 0 bytes (i.e. all files are empty, probably being used just as placeholders) is interrupted during download at a client, then it gets registered as 100% with ActiveEfficiency even though not all the files were downloaded.|
|Package is being transferred to inhibited subnet via SSD||When a Subnet is marked as inhibited, the inter-subnet content transfer still happens|
|Pre-caching exceptions occur in the ActiveEfficiency logs||Pre-caching does not happen for Nomad clients that are on devices which are duplicated in the Configuration Manager database, this may occasionally cause exceptions to occur.|
|Download may fail for Pre-caching jobs that are setup for content that is still being distributed to distribution points||Download may fail for Pre-caching jobs that are setup for content that is still being distributed to distribution points. It is always advisable to setup Pre-caching jobs for content that has been successfully distributed to distribution points.|
|Disabling TLS 1.0 on ActiveEfficiency Server or clients breaks pre-caching and other features.|
Disabling TLS 1.0 on ActiveEfficiency Server would cause the following issues:
Resolution: Please install the latest hotfixes for NomadBranch client and Nomad AdminUI Extensions.
Office 365 updates
|Nomad does not support Office update job when the download is from the Internet||Nomad doesn't support downloading office updates from URLs. For example, the Nomad office job handler will not support this link: https://www.microsoft.com/en-us/download/confirmation.aspx?id=245645d7-6595-403d-bf01-a58ad1d19999.|
|Office 365 download does not resume and fails when Nomadbranch service is restarted or when machines resume from a restart or shutdown.|
A single Office 365 installation leads to multiple Nomad jobs which follows following this sequence:
|>NightWatchman may be able to shutdown the machine when an Office 365 job is running.||Office 365 updates comes as a collection of multiple jobs. When a job completes and another is yet to start, there is small time gap. During this gap, In between the job, ByteRange job manager queue is empty. If at this time, "Nomadbranch -canshutdown" is triggered, it will return "TRUE".|
|Corrupt Content is downloaded successfully, transferred to Office 365 cache and update installation succeeds||As Nomad doesn't receive hashes to validate downloaded contents so user can get the corrupted contents downloaded.|
|ClickToRun stuck in Cancelling Job.. and further install request from Software centre does not triggers any Office Job||ClickToRun is part of Microsoft office 365 and sometimes it stuck while cancelling a job. Also, each new job is provided and handed over to Nomad by ClickToRun. So if ClickToRun is stuck then user won't be able to proceed further.|
|While Running cachecleaner command, user faces "Error: CreateInstance-IUIResourceMgr <FAILED> - Class not registered".||It is caused by UIResource.dll getting unregistered. So workaround is to run the following command(Administrator user context): "regsvr32 UIResource.dll"|
|NomadBranch switches "Activateall" and "Activate=(Content ID)" fails to verify Office 365 content in Nomad Cache.||Office 365 Cache is activated on Nomad service start up. So if, Manually copied some (Valid) data in Nomad Cache, Service needs restart to activate data. - Activate cmd does not impacts office cache.|
|Software Center displays Office 365 Update being downloading even after Nomadbranch Service is stopped (License Expired).||When Nomad license expires, its services is stopped. So in case a download is in progress or being started and nomad notices that its license is expired then it will stop its service. As service is stopped so no status can be returned to Software center and hence it displays it Downloading.|
|Nomad Office Job keeps on going even after reserve space got full.||This is happening in a particular scenario and is related to the case when a full file is being downloaded by Nomad office 365 job. As for full file download, the information about content size is not with Nomad so it keeps on downloading.|
|Nomad Cache refused to clear cache on Peer when other client downloading it from P2P.||The behavior is that the ByteRanges content might or might not get deleted depending on whether it's content is being accessed at that point in time or not. If it does get deleted, the Peer move next available peer and if not then move to DP for downloading.|
|Intermittent issue: Non office job is not cancelling with command line properly.||This issue is a randomly occurring one and sometimes when user tries to cancel a job using nomad command line(Nomadbranch -cancel) then sometimes(randomly) the job is not cancelled. User can rerun the command or in rare case restart the nomadbranch service and then rerun the Cancellation command.|
|Office 365 Update might fail if during download of Office 365 update SMSAgent service is restarted.||Office 365 Update might fail if during download of Office 365 update SMSAgent service is restarted. When download of failed job is re-initiated from Software Center the job download fails. Again ClickToRun being stuck in cancelling is causing this issue.|
|Office 365 Update will sometimes fail if during download of Office 365 update requests a file which does not exist or folder reference does not exist.||This issue is specific to "Office 365 Client Update - Current channel(1609-4)" update. Here it is observed that wrong path is being generated in manifest file from Office 365 which results in update failure due to resources not found. All updates prior to this update were generating correct content path in the manifest file but this specific(Latest update by the time of writing this information) the content path is wrong in manifest file. Also this issue is not limited to Nomad alone but even the CM software center triggered and CM managed update download is also failing.|
|Nomad failed to delete cache with Percent Available disk space value.||This issue has two parts to it:|
|Nomad office updates doesn't support failover to BITS.||Nomad is not supporting "Failover to BITS" scenario in this release.|
The Nomad Dashboard
|Export intermittently generates an Excel file format even if PDF or Word is chosen.||This is random occurrence, but for most of the time the correct export file format is generated.|
|Client Version Distribution Tile fails to load data when expanded||This occurs when extremely large databases encountered.|
|Filter fails to list data when applied on Nomad Dashboard tiles when user changes the size of the opened application (CM Console/Supported Browsers)||Resizing of the console/browser window results in records listed getting empty in UI.|
|Dashboard caching metrics of latest deployment shows 100% download on deployment only while no activity was performed on client side.||Dashboard in this case is actually showing last download detail and not the current one. This is observed for Software Updates and Applications.|
|No In progress report of deployment-due to NULL subnet value.||When there is no subnet value present, Dashboard is not able to show "In Progress" report in chart: "download progress: latest deployment".|
|Dashboard Chart showing incorrect portions when a machine has cancelled and restarted a job.||This issue is caused by another issue where Dashboard caching metrics of latest deployment shows 100% download on deployment.|
|Dashboard is showing data older than the configured one.||Dashboard is showing data older than the configured one. For example if user configured dashboard to show data for 30 days only, which is also a default number but dashboard will be showing data for older than 30 days.|
|Client does not participate in PBA Polling when two clients with the same name exist in two different domains||If a client machine with same hostname name exists in two different domains, during PBA Polling the client does not participitatie|
|Peer replying first in election is elected as Master irrespective of start-up time|
When None of the replying peer is current Master for Given Package and all the replying peers have same weight-age, also all the peers have 100% content for the given package.
During download for the given package Elections is triggered and Peer replying first in election is elected as Master irrespective of start-up time of Nomad service
|Nomad shows incorrect logging indicating connection limit is over.||When there network disconnection during P2P download over http(s), sometimes incorrect logging appears indicating connection limit is over.|
|Elections don't correctly honour uptime||When election weight is same on all peers and one of the peer has greater uptime than everybody else then Nomad may not elect the machine with better uptime as master.|
|In some cases high election weight peer loses election despite same content percentage||When all the peers have almost same content percent then sometimes peer with higher election weight does not win elections.|
Nomad and Windows Servicing Assistant
Windows Servicing Assistant (WSA) is a feature of the Tachyon Agent and 1E Client.
|WSA deployment may restore data, which was captured by a different user||When Replace capture WSA deployment is executed on source machine multiple times, different Nomad PBA Hosts can be honored to capture USMT data each time. During replace restore, target machine can restore any captured USMT data. Nomad PBA doesn't have a notion of version.|
|TS execution fails as Content is not located in Nomad Cache if Nomad Cache was configured on Custom Drive||If Nomad Cache is configured on the Custom drive then TS execution will fail as content will be searched in the default Nomad location.|
|WSA Wizard might requests USB drive even when there is enough disk space.||Even though there is enough disk space on users workstation WSA Wizard might request for USB drive.|
|PBA POLL / FIND fails during WSA TS execution.||If potential PBA host found during WSA prepareness, goes offline during Task Sequence execution, the TS can fail. Enabling HA can reduce the chances of such failures.|