We are attempting to deploy a real large software package (over 4GB) and find that when deploying the action, the client will download roughly half of the package and then come back with download failed. The client log comes up with a “JobDiskLimited - try again in 10 minutes” for that particular action.
I’ve attempted a few things here such as setting the _BESClient_Download_DownloadsCacheLimitMB client setting to 6200 and no go. Tried on multiple machines and platforms (Mac and PC) and find similar results.
Am I missing something here? Is there a default limit to the download cache and if so how do you go about changing it?
I think you are running into limits on how large a single download currently being transfered can be. The DownloadCacheLimitMB adjusts how large the cache for all downloads are but I think you are hitting limits on the size of the currently active download.
Try these settings to raise the default limits depending on whether you are using pre-cache or not:
_BESClient_Download_NormalStageDiskLimitMB (Default is 2048)
_BESClient_Download_PreCacheStageDiskLimitMB (Default is 250)
We’re beginning to see this error come up with some of our actions that use the prefetch command for downloads. We do not have “Start client downloads before constraints are satisfied” checked in our actions.
Could you shed any more light on these settings? I can’t seem to find any documentation on them. Are they new in version 8? Also, does it prevent baselines from continuing when it occurs? It seems like it causes baselines to choke…
In 8.0 we had to redo some of our agent download manager to handle the pre-caching for that checkbox and we introduced some new settings. The two that Tyler mentioned are the most important:
_BESClient_Download_PreCacheStageDiskLimitMB (Default is 250)
_BESClient_Download_NormalStageDiskLimitMB (Default is 2048)
The first setting (PreCache) controls how much data the agent will download when you check the box “Start client downloads before constraints are satisfied”.
The second setting (Normal) will affect all downloads (even if you don’t use that setting). We introduced this setting in 8.0 because we had removed a restriction that prohibited files >2GB from being downloaded so we wanted to add a setting that would allow customers to control the max download size.
I am not sure why this would affect baselines…
If you are hitting this issue, you can just raise the _BESClient_Download_NormalStageDiskLimitMB setting to something higher (maybe 10000?)
Excellent, thanks for the clarification Ben. We’ll make both of those settings higher.
I’m going to look into the baseline “choking” issue a bit more. It looks like it may not be related to the download limits. Yesterday, two baselines were running on a workstation at the same time, and one of them triggered a reboot while the other was still running. The baseline that was still running has not run since, but is supposed to re-run at least 3 times and continue running if an action fails. Restarting the BES client and the system didn’t help.
Terribly sorry to bring this up from the grave. I too am having a very similar issue when trying to perform an action on a client. The logs show the same “JobDiskLimited - try again in 10 minutes” error message.
The responses above make great sense if only I can find out where to change these settings at. I feel like I have looked everywhere but it will likely be something simple. Any advice?
Edit:
After going through a stage of panic, I opened my eyes a bit wider and discovered the following statement hidden in the “View Action Info” window.
The download size exceeds the maximum value set in the client setting _BESClient_Download_PreCacheStageDiskLimitMB, which can be modified through the Edit Computer Settings dialog.
Hopefully this may help someone else who finds themselves stuck also.
OK, seem to up against a brick wall at the moment. Trying to deply Office 2010 SP1. Using the Windows Software Distribution Wizard I created 2 tasks. One is the office files. Pointed the wizard at the folder containing the extracted CD. It creates a task and a .tmp file of approx. 1 GB. The second is a single .msp config file to be using during the office install. I then then combine the two download/continue/extract statements into a third task that is pushed to the users.
Everything seems to be going OK until the extract. I deploy, using a message (not using the start client downloads before constraints…) and it downloads the two files OK. When I click on “Take Action” on the message it copies the two files into the “actionsite__Download” folder but never extracts them.
Command succeeded (evaluated true) continue if {(size of it = 5450002 AND sha1 of it = “0cde6653c537dde8ebd1e9a206e9b79c457a67b7”) of file “IKO_SP1_Default_AllAppsMSP.tmp” of folder “__Download”} (fixlet 8230)
Command failed (Relevance substitution failed) continue if {(size of it = 1056859616 AND sha1 of it = “792b04d89a48510e177f78e263b365eed2e0885c”) of file “Office2010wSP1.tmp” of folder “__Download”} (fixlet 8230)
It appears it downloaded and extracted the one file OK but failed on the large 1GB file. It’s almost like there is a mis-match on the size or SHA1.
OK, here’s an odd one. I re-pushed the same task without any user interaction and it installed successfully??!!
To verify this I rolled back my VM and re-pushed the same task display a message to the user and having them click “Take Action” and the same error as above. (.TMP files copied to actionsite__Download but not extracted and action fails!) Pushed the task again having NO user interaction and it installs no problem! VERY WEIRD. Can anyone provide some insight here??
FYI. I have set the PreCacheStageDiskLimitMB to 2048 and the NormalStageDiskLimitMB to 3072