Once upon a time, it was so simple. Six questions with yes or no, and you can just put in place UPM following the docs, and performing a simple if-then-X-Y-set-so the sequence.
But then came PVD and VDI -in-a-Box, and people began running provisioned XenApp on XenServer, and goodness-knows-what-else, and suddenly he had to think very seriously some aspects of UPM configuration. Not because you might break UPM (I mean, could force UPM to lose data), no. But it was possible to create configurations that have been sub-optimal in terms of connection / disconnection speed, or were too expensive in terms of resources.
A sea of troubles?
The goals of the Profile Management (the concept, rather than the code) is to make your profile as soon as possible, and keep it in a safe place, even when you are not using it . Share all of your sessions is also important
The techniques we use to achieve these goals are :.
- DC Profile
- the local caching profiles
- Background Fill
- active Write Back
continuous profile leads to rapid logons, eliminating much of the network copy. However, UPM has yet to analyze network file and folder structure, to create a local copy using Reparse Points to act as placeholders for the actual files. Again, generally accelerated, but there are some drawbacks:
- If the profile contains many small files, the profile continuously can not save bandwidth much network compared to copying files
- If local profiles are maintained on shared storage, you must be aware of IOPS incurred by the implementation of the Reparse Points - even if the physical space required is minimal, depending on your storage design, up to 6 I / O operations may be required by Reparse point, which can be as costly as copying the file, but worse in terms of the impact on other users, such as I / O is "bursty"
- If the profile was only partially cover -filled by the end of the session, the locally cached profile will be saved with the analysis points, which should be deleted and recreated on the next logon, which causes more IOPS.
Although the streaming profile can speed connections, the technique can have the greatest impact on the charging time profile is local caching of profiles, because if the profile is already (mostly) local, then the loading cost, it is quite small. We still need to launch an audit of updates to other sessions, but that is usually a quick operation. So the local cache profile layout is good what to do, but there are some drawbacks:
- On XenApp and Terminal Services (RDS) environments, it is very easy consume a lot of disk space with "stale" profiles that are rarely (or never) accessible
- in a provisioned environment, cached profiles will not survive a power cycle - but the rise is planned that frequent reboots are solving the problem of disk space [1945007!]
- full profile must be cached, which means disabling a continuous profile to be sure you have all the profile. The first login will be much longer - giving a bad first impression of the technology
Background Fill (aka Always Cache) tries to get the best of both worlds !. If you want to speed first Streaming profile connection, but the long term benefits of having a local copy, the solution may be to enable Always Cache. The profile is distributed at the first login, but UPM background continues to run, hide and fill all points of analysis (the "placeholders") until the entire profile copied. Then, on the next logon, there is a locally cached profile in place, and the login is fast. There are unfortunately some drawbacks:
- The background downloading files consumes network bandwidth, and (in pathological cases) can delay copying the files you want urgent
- If the. profile has not been completely downloaded when the session ends (which will be true if the session is unusually short, or your profile is exceptionally high), the profile will be saved with the analysis points, which will be recreated on the next logon because they can not be reused by the new session.
active Write is a feature designed to help secure settings changes in environments (provisioned) Volatile, including VDI-in-a box. The changes are copied to the network profile updated when files are closed, subject to any update in a 10-second window after closing. It is useful if users tend to disconnect, not logoff and then complain that their changes are lost if there is a power failure. And disadvantages:
- Only files are saved. So if applications rely on file and registry changes matched the profile can be left in an inconsistent state. A better solution is today used for PVD.
- In order to minimize network traffic resulting from many updates, UPM applies a "quiet time" 5 minutes after each burst of AWB traffic when updates are waiting to be sent. If the user logs out during this period, the disconnection may be slightly delayed.
And thus the native hue of resolution ...
We'll start with the question of whether or not to cache profiles.
profiles Do not remove cached locally on logoff if ...
- The (virtual) XenDesktop machine is volatile and will be destroyed on logoff. Deleting profiles at logoff costs time. If the machine is about to be destroyed, why?
- The profile is stored in a personal vDisk. PVD is designed to store profiles in the vDisk, so UPM profile should leave in place for developing countries to manage
- A special case: VDI-in-a-Box is in use. There are two scenarios here. In the first scenario, VDI-in-a-box is deployed with PVD, to treat as above. In the second scenario, PVD is not present; logoff on the machine will be destroyed anyway, so why waste time deleting the profile?
- The machine is "assigned" to a user (XenDesktop) and by implication is persistent. The fastest connections will most likely be achieved by caching the profile on a disk (persistent) space.
- The machine is dedicated to the use of a small number of users (XenApp / TS / RDS), and a suitably large persistent disk. The case here is a small departmental server, where the number of users is limited.
- The availability of the network. If the network is limited availability, it goes back to the / mobile in the static case Sextuple Way. Configure Offline profiles, which forces locally cached profiles to remember and disables the continuous profile.
profiles You must remove cached locally on logoff if ...
- The machine is a persistent server (XenApp / TS / RDS). Delete profiles on logoff, to prevent the proliferation of outdated profiles. You can also use delprof or similar to remove obsolete profiles that are not used for "long"
- The machine is pooled XenDesktop. Delete profile on logoff if the workstations can be recycled between users, rather than being created / destroyed at the request
That said, it is always "safe" to delete the profiles locally cached on logoff -. Unlike Microsoft roaming profile, UPM conceptually the profile is always and only in its entirety on the magazine of network users, making brief forays to cast its shadow on the parameters. (In contrast, Microsoft's roaming profile exists in quantum as many terminals simultaneously, but only one of them turns out to be true - the last to be "seen" by a disconnection appears to have been the "real" d one throughout.)
Thus, having made our decision to cache or not to cache the local profile, which helps us decide how to define the continuous profile.
In principle, if we cover the profile, we really want to get the whole low profile before logging off, so there are two choices
1 choice :. Disable the streaming profile. This ensures that the profile will be present in full upon logon ends. Disadvantages :.
- This means that the first logon will slow, while the complete profile is copied down
Option 2: Enable Streaming profile but also to always configure cache and a size 0. this causes a background thread to copy the low profile, file by file, until the complete profile is copied down. Disadvantages:
- There are no guarantees that the copy is completed when the user logs off, in which case UPM will record a copy locally cached profile with Reparse Points - ceux- these have to be re-created by the next session, as (for security reasons), they can not be reused by the next session. UPM will eventually catch up, and all files will eventually be cached locally, so the cost is (only) the additional IOPS to recreate the analysis points.
- There is no way to inform the user that the copy is complete
- downloading files in the background, such as large video files, could delay the access files that the user wants urgently. pathological case - see below
Ultimately this choice for the administrator if the UPM Setup verification tool is set to recommend Option 1 -. to disable the streaming profile
The other situation when the profiles are not cached locally, usually occurs when there are no persistent store, in which case to allow streaming profile makes perfect sense. Disadvantages :.
- The IOPS consideration - each logon will be accompanied by a burst of creative analysis point, which can impose a heavy burden if the local profile cache resides on shared storage
So now we turn to still Cache, which is designed to work alongside a continuous profile. There are two scenarios where you can set Always Cache
Scenario 1 :. You want both a fast connection and also for the entire low profile, because you have decided that you want to keep the local cached profiles on logoff. Always set Cache, with a size of 0 (zero)
Scenario 2 :. You want a fast connection, but also want to upload larger files in the background, so you should not expect if you decide to use them. Always set Cache, with the format set at the lower limit of the size of the files you want to automatically populate the cache, in MB
Large files, however, are usually "business data" rather the parameters, and are more likely to be found in My Documents, My pictures or My videos. These records are best kept separate parts of the profile parameters, so that the folder redirection is a better alternative.
UPM Config check will always recommend a value of 0 (zero) if delete profiles locally cached on logoff is set, otherwise it will simply report the value set, but will make no recommendation.
Finally, we look at active Write Back. This is designed to write new file changes on the network to protect against data loss in unstable environments.
If PVD is active, disable Active Write Back.
If the machine is volatile, disable active Write Back.
Otherwise, activate active Write Back. Disadvantages:
- Assets Write Back does not capture registry changes. Applications may require two registry changes and file changes in order to retain certain functions to operate.
UPM Config Check PVD recommend the use in environments of volatile workstations. In environments of volatile servers, UPM Config Check recommend active Write Back with a warning.
tall Businesses and time
Also known as the conclusion.
This article replaces the advice to set the four political
- continuous profile
- Delete Profiles cached locally on logoff
- always Cache
- active Write Back
presented in eDocs at http://support.citrix.com/proddocs/topic/user-profile-manager-sou/upm-plan-decide- wrapper.html, taking into account the presence of
- provisioning technologies such as PVS and MCS
- PVD
- VDI-in-a-box
the name of the action ...
A new version of the UpmConfigCheck tool that implements the recommendations above (and many other improvements) is in preparation (it is coded, but still being tested). I'll post again when it is available.
0 Komentar