Failed to find file storage for storage type "passivation:"
Hi,
When I start my pega jboss node, i get an error message about a passivation file that it can't find.
The message :
Hi,
When I start my pega jboss node, i get an error message about a passivation file that it can't find.
The message :
2020-01-07 10:25:11,078 INFO [stdout] (Log4j2-TF-7-AsyncLoggerConfig-1) 2020-01-07 10:25:11,078 [ default task-1] [ ] [ ] [ ] (les.storage.FileStorageManager) ERROR slf04059.n18.an.cnav| Proprietary information hidden - Failed to find file storage for storage type "passivation:" when looking up file specification "file://passivation://PassivationData//A//HU0VV0QGWLUHG7UUV9H7T3UT3JLDIUIKEA//STANDARD//STANDARD-pzThread.ser"; FileStorageManager may not having been initialized correctly.
2020-01-07 10:25:11,106 INFO [stdout] (Log4j2-TF-7-AsyncLoggerConfig-1) 2020-01-07 10:25:11,079 [ default task-1] [ ] [ ] [ ] (ngineinterface.service.HttpAPI) ERROR slf04059.n18.an.cnav| Proprietary information hidden - : java.lang.NullPointerException
2020-01-07 10:25:11,106 INFO [stdout] (Log4j2-TF-7-AsyncLoggerConfig-1) java.lang.NullPointerException: null
2020-01-07 10:25:11,106 INFO [stdout] (Log4j2-TF-7-AsyncLoggerConfig-1) at com.pega.pegarules.pub.util.PRFile.exists(PRFile.java:137) ~[prpublic.jar:?]
2020-01-07 10:25:11,106 INFO [stdout] (Log4j2-TF-7-AsyncLoggerConfig-1) at com.pega.pegarules.session.internal.mgmt.base.FilesystemStorage$RegularIO.readObjectFromInputStream(FilesystemStorage.java:991) ~[prprivate.jar:?]
2020-01-07 10:25:11,106 INFO [stdout] (Log4j2-TF-7-AsyncLoggerConfig-1) at com.pega.pegarules.session.internal.mgmt.base.FilesystemStorage$RegularIO.readFromDisk(FilesystemStorage.java:986) ~[prprivate.jar:?]
2020-01-07 10:25:11,106 INFO [stdout] (Log4j2-TF-7-AsyncLoggerConfig-1) at com.pega.pegarules.session.internal.mgmt.base.FilesystemStorage.activateThread(FilesystemStorage.java:425) ~[prprivate.jar:?]
2020-01-07 10:25:11,106 INFO [stdout] (Log4j2-TF-7-AsyncLoggerConfig-1) at com.pega.pegarules.session.internal.mgmt.base.NodeRequestorMgt.readThread(NodeRequestorMgt.java:1004) ~[prprivate.jar:?]
2020-01-07 10:25:11,106 INFO [stdout] (Log4j2-TF-7-AsyncLoggerConfig-1) at com.pega.pegarules.session.internal.mgmt.base.RequestorThreadSync.getOrCreateThread(RequestorThreadSync.java:178) ~[prprivate.jar:?]
2020-01-07 10:25:11,106 INFO [stdout] (Log4j2-TF-7-AsyncLoggerConfig-1) at com.pega.pegarules.session.internal.mgmt.base.RequestorThreadSync.getOrCreateThread(RequestorThreadSync.java:164) ~[prprivate.jar:?]
2020-01-07 10:25:11,106 INFO [stdout] (Log4j2-TF-7-AsyncLoggerConfig-1) at com.pega.pegarules.session.internal.mgmt.base.RequestorThreadSync.getOrCreateThread(RequestorThreadSync.java:160) ~[prprivate.jar:?]
2020-01-07 10:25:11,106 INFO [stdout] (Log4j2-TF-7-AsyncLoggerConfig-1) at com.pega.pegarules.session.internal.mgmt.base.RequestorPassivation.configureRequestorImpl(RequestorPassivation.java:502) ~[prprivate.jar:?]
2020-01-07 10:25:11,106 INFO [stdout] (Log4j2-TF-7-AsyncLoggerConfig-1) at com.pega.pegarules.session.internal.mgmt.PRRequestorBase.configureRequestor(PRRequestorBase.java:485) ~[prprivate.jar:?]
2020-01-07 10:25:11,106 INFO [stdout] (Log4j2-TF-7-AsyncLoggerConfig-1) at com.pega.pegarules.session.internal.mgmt.PRRequestorImpl.<init>(PRRequestorImpl.java:330) ~[prprivate.jar:?]
The end of my prconfig.xml :
<env name="session/ha/quiesce/customSessionInvalidationMethod" value="configurable-cookie-invalidator" />
<env name="session/ha/quiesce/cookieToInvalidate" value="BIGipServerP_VL994_RET_FORM_8080_1" />
<!-- Work recovery related settings, the EnableUserNotification could be changed using HA lading page(Cluster Settings page) -->
<env name="session/ha/crash/RecordWorkInProgress" value="true"/>
<env name="session/ha/Enabled" value="true"/>
<env name="session/ha/crash/EnableUserNotification" value="true"/>
<env name="session/ha/quiesce/strategy" value="immediateDrain"/>
<!-- Shared Passivation, Should be common for all the nodes in a cluster -->
<env name="initialization/persistrequestor/storage" value="filesystem" />
<env name="storage/class/passivation:/rootpath" value="/pegadata/sessions" />
</pegarules>
/pegadata/sessions exists on the machine and has the adequate access rights (i can create a folder with the user that is running my pega jboss).
Any idea what else to look for ?
Best regards
Accepted Solution
CARSAT-RA
FR
Hi,
we use autofs to map the network share for the passivation.
when not used, autofs unmounts the share (and the access rights are set to root. when accessed, after a moment, the share is mounted by autofs and the access rights are changed to the pega/jboss system user.
pega doesn't do retry ; this makes our error
=> our problem was because the network share used for passivation takes too long to be mapped, and pega doesn't retry.
=> a solution would be tu use a more reliable way of mapping the share than autofs,
or
find a parameter in pega to do the retry
or
find a parameter in autofs to pre map the share
or
as you say, to forget the network shares for passivation and do it in db
i guess we will try the passivation in db
Thanks everybody
CARSAT-RA
FR
Hi,
Thanks for this quick reaction.
I have pega 8.2.
So with this version of pega it's not possible to use a filesystem passivation ? or it's just that the default way of doing passivation for this version of pega is database but we can change it ?
We assumed that this parameter in prconfig.xml was the right one to use filesystem :
<env name="initialization/persistrequestor/storage" value="filesystem" />
where we wrong ?
Pegasystems Inc.
US
You can still use filesystem passivation, but I think the question we have is why do you want to do that instead of using the database?
This help article may aid in clearing things up: https://community.pega.com/sites/default/files/help_v83/procomhelpmain.htm#engine/high-availability/architecture/eng-high-availability-shared-storage-con.htm
Is the location being specified available to all servers in the cluster?
Pegasystems Inc.
US
You should NOT use filesystem as the performance is not as good for your version (not to mention maintenance issues with filesystem). Is there any reason you do not want to use database as passivation storage type? For 8.2, you do not have to specify the prconfig setting as database will be used since HA is enabled.
CARSAT-RA
FR
Hi,
Thanks for this piece of information.
The reason is somebody from pega.com came to help us in june last year and said that for our version it was a good practice tu use the filesystem...
So for our version of pega, for performance reasons, we should remove the following block from our prconfig.xml file and leave the rest ; am i right ?
<!-- Shared Passivation, Should be common for all the nodes in a cluster -->
<env name="initialization/persistrequestor/storage" value="filesystem" />
<env name="storage/class/passivation:/rootpath" value="/pegadata/sessions" />
Accepted Solution
CARSAT-RA
FR
Hi,
we use autofs to map the network share for the passivation.
when not used, autofs unmounts the share (and the access rights are set to root. when accessed, after a moment, the share is mounted by autofs and the access rights are changed to the pega/jboss system user.
pega doesn't do retry ; this makes our error
=> our problem was because the network share used for passivation takes too long to be mapped, and pega doesn't retry.
=> a solution would be tu use a more reliable way of mapping the share than autofs,
or
find a parameter in pega to do the retry
or
find a parameter in autofs to pre map the share
or
as you say, to forget the network shares for passivation and do it in db
i guess we will try the passivation in db
Thanks everybody